Publications by authors named "Issam El Naqa"

188 Publications

Investigating the SPECT Dose-Function Metrics Associated With Radiation-Induced Lung Toxicity Risk in Patients With Non-small Cell Lung Cancer Undergoing Radiation Therapy.

Adv Radiat Oncol 2021 May-Jun;6(3):100666. Epub 2021 Feb 7.

Department of Radiation Oncology, University of Michigan, Ann Arbor, Michigan.

Purpose: Dose to normal lung has commonly been linked with radiation-induced lung toxicity (RILT) risk, but incorporating functional lung metrics in treatment planning may help further optimize dose delivery and reduce RILT incidence. The purpose of this study was to investigate the impact of the dose delivered to functional lung regions by analyzing perfusion (Q), ventilation (V), and combined V/Q single-photon-emission computed tomography (SPECT) dose-function metrics with regard to RILT risk in patients with non-small cell lung cancer (NSCLC) patients who received radiation therapy (RT).

Methods And Materials: SPECT images acquired from 88 patients with locally advanced NSCLC before undergoing conventionally fractionated RT were retrospectively analyzed. Dose was converted to the nominal dose equivalent per 2 Gy fraction, and SPECT intensities were normalized. Regional lung segments were defined, and the average dose delivered to each lung region was quantified. Three functional categorizations were defined to represent low-, normal-, and high-functioning lungs. The percent of functional lung category receiving ≥20 Gy and mean functional intensity receiving ≥20 Gy (iV) were calculated. RILT was defined as grade 2+ radiation pneumonitis and/or clinical radiation fibrosis. A logistic regression was used to evaluate the association between dose-function metrics and risk of RILT.

Results: By analyzing V/Q normalized intensities and functional distributions across the population, a wide range in functional capability (especially in the ipsilateral lung) was observed in patients with NSCLC before RT. Through multivariable regression models, global lung average dose to the lower lung was found to be significantly associated with RILT, and Q and V iV were correlated with RILT when using ipsilateral lung metrics. Through a receiver operating characteristic analysis, combined V/Q low-function receiving ≥20 Gy (low-functioning V/Q) in the ipsilateral lung was found to be the best predictor (area under the curce: 0.79) of RILT risk.

Conclusions: Irradiation of the inferior lung appears to be a locational sensitivity for RILT risk. The multivariable correlation between ipsilateral lung iV and RILT, as well as the association of low-functioning V/Q and RILT, suggest that irradiating low-functioning regions in the lung may lead to higher toxicity rates.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1016/j.adro.2021.100666DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC8010578PMC
February 2021

Neurocognitive Effects and Necrosis in Childhood Cancer Survivors Treated With Radiation Therapy: A PENTEC Comprehensive Review.

Int J Radiat Oncol Biol Phys 2021 Mar 9. Epub 2021 Mar 9.

Medical Physics Department, Centro di Riferimento Oncologico di Aviano (CRO) IRCCS, Aviano, Italy.

Purpose: A PENTEC review of childhood cancer survivors who received brain radiation therapy (RT) was performed to develop models that aid in developing dose constraints for RT-associated central nervous system (CNS) morbidities.

Methods And Materials: A comprehensive literature search, through the PENTEC initiative, was performed to identify published data pertaining to 6 specific CNS toxicities in children treated with brain RT. Treatment and outcome data on survivors were extracted and used to generate normal tissue complication probability (NTCP) models.

Results: The search identified investigations pertaining to 2 of the 6 predefined CNS outcomes: neurocognition and brain necrosis. For neurocognition, models for 2 post-RT outcomes were developed to (1) calculate the risk for a below-average intelligence quotient (IQ) (IQ <85) and (2) estimate the expected IQ value. The models suggest that there is a 5% risk of a subsequent IQ <85 when 10%, 20%, 50%, or 100% of the brain is irradiated to 35.7, 29.1, 22.2, or 18.1 Gy, respectively (all at 2 Gy/fraction and without methotrexate). Methotrexate (MTX) increased the risk for an IQ <85 similar to a generalized uniform brain dose of 5.9 Gy. The model for predicting expected IQ also includes the effect of dose, age, and MTX. Each of these factors has an independent, but probably cumulative effect on IQ. The necrosis model estimates a 5% risk of necrosis for children after 58.9 Gy or 59.9 Gy (2 Gy/fraction) to any part of the brain if delivered as primary RT or reirradiation, respectively.

Conclusions: This PENTEC comprehensive review establishes objective relationships between patient age, RT dose, RT volume, and MTX to subsequent risks of neurocognitive injury and necrosis. A lack of consistent RT data and outcome reporting in the published literature hindered investigation of the other predefined CNS morbidity endpoints.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1016/j.ijrobp.2020.11.073DOI Listing
March 2021

Quantitative Molecular Positron Emission Tomography Imaging Using Advanced Deep Learning Techniques.

Annu Rev Biomed Eng 2021 Apr 2. Epub 2021 Apr 2.

Department of Machine Learning, Moffitt Cancer Center, Tampa, Florida 33612, USA.

The widespread availability of high-performance computing and the popularity of artificial intelligence (AI) with machine learning and deep learning (ML/DL) algorithms at the helm have stimulated the development of many applications involving the use of AI-based techniques in molecular imaging research. Applications reported in the literature encompass various areas, including innovative design concepts in positron emission tomography (PET) instrumentation, quantitative image reconstruction and analysis techniques, computer-aided detection and diagnosis, as well as modeling and prediction of outcomes. This review reflects the tremendous interest in quantitative molecular imaging using ML/DL techniques during the past decade, ranging from the basic principles of ML/DL techniques to the various steps required for obtaining quantitatively accurate PET data, including algorithms used to denoise or correct for physical degrading factors as well as to quantify tracer uptake and metabolic tumor volume for treatment monitoring or radiation therapy treatment planning and response prediction. This review also addresses future opportunities and current challenges facing the adoption of ML/DL approaches and their role in multimodality imaging. Expected final online publication date for the , Volume 23 is June 2021. Please see http://www.annualreviews.org/page/journal/pubdates for revised estimates.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1146/annurev-bioeng-082420-020343DOI Listing
April 2021

Radiomic and radiogenomic modeling for radiotherapy: strategies, pitfalls, and challenges.

J Med Imaging (Bellingham) 2021 May 23;8(3):031902. Epub 2021 Mar 23.

Moffitt Cancer Center and Research Institute, Department of Machine Learning, Tampa, Florida, United States.

The power of predictive modeling for radiotherapy outcomes has historically been limited by an inability to adequately capture patient-specific variabilities; however, next-generation platforms together with imaging technologies and powerful bioinformatic tools have facilitated strategies and provided optimism. Integrating clinical, biological, imaging, and treatment-specific data for more accurate prediction of tumor control probabilities or risk of radiation-induced side effects are high-dimensional problems whose solutions could have widespread benefits to a diverse patient population-we discuss technical approaches toward this objective. Increasing interest in the above is specifically reflected by the emergence of two nascent fields, which are distinct but complementary: radiogenomics, which broadly seeks to integrate biological risk factors together with treatment and diagnostic information to generate individualized patient risk profiles, and radiomics, which further leverages large-scale imaging correlates and extracted features for the same purpose. We review classical analytical and data-driven approaches for outcomes prediction that serve as antecedents to both radiomic and radiogenomic strategies. Discussion then focuses on uses of conventional and deep machine learning in radiomics. We further consider promising strategies for the harmonization of high-dimensional, heterogeneous multiomics datasets (panomics) and techniques for nonparametric validation of best-fit models. Strategies to overcome common pitfalls that are unique to data-intensive radiomics are also discussed.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1117/1.JMI.8.3.031902DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7985651PMC
May 2021

Requirements and reliability of AI in the medical context.

Phys Med 2021 Mar 13;83:72-78. Epub 2021 Mar 13.

Department of Machine Learning, H. Lee. Moffitt Cancer Center, Tampa, FL, USA. Electronic address:

The digital information age has been a catalyst in creating a renewed interest in Artificial Intelligence (AI) approaches, especially the subclass of computer algorithms that are popularly grouped into Machine Learning (ML). These methods have allowed one to go beyond limited human cognitive ability into understanding the complexity in the high dimensional data. Medical sciences have seen a steady use of these methods but have been slow in adoption to improve patient care. There are some significant impediments that have diluted this effort, which include availability of curated diverse data sets for model building, reliable human-level interpretation of these models, and reliable reproducibility of these methods for routine clinical use. Each of these aspects has several limiting conditions that need to be balanced out, considering the data/model building efforts, clinical implementation, integration cost to translational effort with minimal patient level harm, which may directly impact future clinical adoption. In this review paper, we will assess each aspect of the problem in the context of reliable use of the ML methods in oncology, as a representative study case, with the goal to safeguard utility and improve patient care in medicine in general.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1016/j.ejmp.2021.02.024DOI Listing
March 2021

A deep survival interpretable radiomics model of hepatocellular carcinoma patients.

Phys Med 2021 Feb 10;82:295-305. Epub 2021 Mar 10.

Department of Machine Learning, Moffitt Cancer Center, Tampa, FL, USA.

This work aims to identify a new radiomics signature using imaging phenotypes and clinical variables for risk prediction of overall survival (OS) in hepatocellular carcinoma (HCC) patients treated with stereotactic body radiation therapy (SBRT). 167 patients were retrospectively analyzed with repeated nested cross-validation to mitigate overfitting issues. 56 radiomic features were extracted from pre-treatment contrast-enhanced (CE) CT images. 37 clinical factors were obtained from patients' electronic records. Variational autoencoders (VAE) based survival models were designed for radiomics and clinical features and a convolutional neural network (CNN) survival model was used for the CECT. Finally, radiomics, clinical and raw image deep learning network (DNN) models were combined to predict the risk probability for OS. The final models yielded c-indices of 0.579 (95%CI: 0.544-0.621), 0.629 (95%CI: 0.601-0.643), 0.581 (95%CI: 0.553-0.613) and 0.650 (95%CI: 0.635-0.683) for radiomics, clinical, image input and combined models on nested cross validation scheme, respectively. Integrated gradients method was used to interpret the trained models. Our interpretability analysis of the DNN showed that the top ranked features were clinical liver function and liver exclusive of tumor radiomics features, which suggests a prominent role of side effects and toxicities in liver outside the tumor region in determining the survival rate of these patients. In summary, novel deep radiomic analysis provides improved performance for risk assessment of HCC prognosis compared with Cox survival models and may facilitate stratification of HCC patients and personalization of their treatment strategies. Liver function was found to contribute most to the OS for these HCC patients and radiomics can aid in their management.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1016/j.ejmp.2021.02.013DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC8035300PMC
February 2021

Integrating Multiomics Information in Deep Learning Architectures for Joint Actuarial Outcome Prediction in Non-Small Cell Lung Cancer Patients After Radiation Therapy.

Int J Radiat Oncol Biol Phys 2021 Feb 1. Epub 2021 Feb 1.

Department of Radiation Oncology, University of Michigan, Ann Arbor, Michigan.

Purpose: Novel actuarial deep learning neural network (ADNN) architectures are proposed for joint prediction of radiation therapy outcomes-radiation pneumonitis (RP) and local control (LC)-in stage III non-small cell lung cancer (NSCLC) patients. Unlike normal tissue complication probability/tumor control probability models that use dosimetric information solely, our proposed models consider complex interactions among multiomics information including positron emission tomography (PET) radiomics, cytokines, and miRNAs. Additional time-to-event information is also used in the actuarial prediction.

Methods And Materials: Three architectures were investigated: ADNN-DVH considered dosimetric information only; ADNN-com integrated multiomics information; and ADNN-com-joint combined RP2 (RP grade ≥2) and LC prediction. In these architectures, differential dose-volume histograms (DVHs) were fed into 1D convolutional neural networks (CNN) for extracting reduced representations. Variational encoders were used to learn representations of imaging and biological data. Reduced representations were fed into Surv-Nets to predict time-to-event probabilities for RP2 and LC independently and jointly by incorporating time information into designated loss functions.

Results: Models were evaluated on 117 retrospective patients and were independently tested on 25 newly accrued patients prospectively. A multi-institutional RTOG0617 data set of 327 patients was used for external validation. ADNN-DVH yielded cross-validated c-indexes (95% confidence intervals) of 0.660 (0.630-0.690) for RP2 prediction and 0.727 (0.700-0.753) for LC prediction, outperforming a generalized Lyman model for RP2 (0.613 [0.583-0.643]) and a generalized log-logistic model for LC (0.569 [0.545-0.594]). The independent internal test and external validation yielded similar results. ADNN-com achieved an even better performance than ADNN-DVH on both cross-validation and independent internal test. Furthermore, ADNN-com-joint, which yielded performance similar to ADNN-com, realized joint prediction with c-indexes of 0.705 (0.676-0.734) for RP2 and 0.740 (0.714-0.765) for LC and achieved an area under a free-response receiving operator characteristic curve (AU-FROC) of 0.729 (0.697-0.773) for the joint prediction of RP2 and LC.

Conclusion: Novel deep learning architectures that integrate multiomics information outperformed traditional normal tissue complication probability/tumor control probability models in actuarial prediction of RP2 and LC.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1016/j.ijrobp.2021.01.042DOI Listing
February 2021

Stereotactic Body Radiation Therapy for Spinal Metastases: Tumor Control Probability Analyses and Recommended Reporting Standards.

Int J Radiat Oncol Biol Phys 2021 Jan 27. Epub 2021 Jan 27.

Department of Radiation Oncology, University of California at Davis, Sacramento, California.

Purpose: We sought to investigate the tumor control probability (TCP) of spinal metastases treated with stereotactic body radiation therapy (SBRT) in 1 to 5 fractions.

Methods And Materials: PubMed-indexed articles from 1995 to 2018 were eligible for data extraction if they contained SBRT dosimetric details correlated with actuarial 2-year local tumor control rates. Logistic dose-response models of collected data were compared in terms of physical dose and 3-fraction equivalent dose.

Results: Data were extracted from 24 articles with 2619 spinal metastases. Physical dose TCP modeling of 2-year local tumor control from the single-fraction data were compared with data from 2 to 5 fractions, resulting in an estimated α/β = 6 Gy, and this was used to pool data. Acknowledging the uncertainty intrinsic to the data extraction and modeling process, the 90% TCP corresponded to 20 Gy in 1 fraction, 28 Gy in 2 fractions, 33 Gy in 3 fractions, and (with extrapolation) 40 Gy in 5 fractions. The estimated TCP for common fractionation schemes was 82% at 18 Gy, 90% for 20 Gy, and 96% for 24 Gy in a single fraction, 82% for 24 Gy in 2 fractions, and 78% for 27 Gy in 3 fractions.

Conclusions: Spinal SBRT with the most common fractionation schemes yields 2-year estimates of local control of 82% to 96%. Given the heterogeneity in the tumor control estimates extracted from the literature, with variability in reporting of dosimetry data and the definition of and statistical methods of reporting tumor control, care should be taken interpreting the resultant model-based estimates. Depending on the clinical intent, the improved TCP with higher dose regimens should be weighed against the potential risks for greater toxicity. We encourage future reports to provide full dosimetric data correlated with tumor local control to allow future efforts of modeling pooled data.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1016/j.ijrobp.2020.11.021DOI Listing
January 2021

Artificial Intelligence for Response Evaluation With PET/CT.

Semin Nucl Med 2021 Mar 11;51(2):157-169. Epub 2020 Nov 11.

Department of Radiation Oncology, Physics Division, University of Michigan, Ann Arbor, MI. Electronic address:

Positron emission tomography (PET)/computed tomography (CT) are nuclear diagnostic imaging modalities that are routinely deployed for cancer staging and monitoring. They hold the advantage of detecting disease related biochemical and physiologic abnormalities in advance of anatomical changes, thus widely used for staging of disease progression, identification of the treatment gross tumor volume, monitoring of disease, as well as prediction of outcomes and personalization of treatment regimens. Among the arsenal of different functional imaging modalities, nuclear imaging has benefited from early adoption of quantitative image analysis starting from simple standard uptake value normalization to more advanced extraction of complex imaging uptake patterns; thanks to application of sophisticated image processing and machine learning algorithms. In this review, we discuss the application of image processing and machine/deep learning techniques to PET/CT imaging with special focus on the oncological radiotherapy domain as a case study and draw examples from our work and others to highlight current status and future potentials.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1053/j.semnuclmed.2020.10.003DOI Listing
March 2021

Liver metastasis restrains immunotherapy efficacy via macrophage-mediated T cell elimination.

Nat Med 2021 01 4;27(1):152-164. Epub 2021 Jan 4.

Department of Surgery, University of Michigan, Ann Arbor, MI, USA.

Metastasis is the primary cause of cancer mortality, and cancer frequently metastasizes to the liver. It is not clear whether liver immune tolerance mechanisms contribute to cancer outcomes. We report that liver metastases diminish immunotherapy efficacy systemically in patients and preclinical models. Patients with liver metastases derive limited benefit from immunotherapy independent of other established biomarkers of response. In multiple mouse models, we show that liver metastases siphon activated CD8 T cells from systemic circulation. Within the liver, activated antigen-specific FasCD8 T cells undergo apoptosis following their interaction with FasLCD11bF4/80 monocyte-derived macrophages. Consequently, liver metastases create a systemic immune desert in preclinical models. Similarly, patients with liver metastases have reduced peripheral T cell numbers and diminished tumoral T cell diversity and function. In preclinical models, liver-directed radiotherapy eliminates immunosuppressive hepatic macrophages, increases hepatic T cell survival and reduces hepatic siphoning of T cells. Thus, liver metastases co-opt host peripheral tolerance mechanisms to cause acquired immunotherapy resistance through CD8 T cell deletion, and the combination of liver-directed radiotherapy and immunotherapy could promote systemic antitumor immunity.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1038/s41591-020-1131-xDOI Listing
January 2021

Tumor Control Probability of Radiosurgery and Fractionated Stereotactic Radiosurgery for Brain Metastases.

Int J Radiat Oncol Biol Phys 2020 Dec 31. Epub 2020 Dec 31.

Department of Radiation Oncology and Molecular Radiation Sciences, Johns Hopkins University School of Medicine, Baltimore, Maryland.

Purpose: As part of the American Association of Physicists in Medicine Working Group on Stereotactic Body Radiotherapy, tumor control probability (TCP) after stereotactic radiosurgery (SRS) and fractionated stereotactic radiosurgery (fSRS) for brain metastases was modeled based on pooled dosimetric and clinical data from published English-language literature.

Methods And Materials: PubMed-indexed studies published between January 1995 and September 2017 were used to evaluate dosimetric and clinical predictors of TCP after SRS or fSRS for brain metastases. Eligible studies had ≥10 patients and included detailed dose-fractionation data with corresponding ≥1-year local control (LC) data, typically evaluated as a >20% increase in diameter of the targeted lesion using the pre-SRS diameter as a reference.

Results: Of 2951 potentially eligible manuscripts, 56 included sufficient dose-volume data for analyses. Accepting that necrosis and pseudoprogression can complicate the assessment of LC, for tumors ≤20 mm, single-fraction doses of 18 and 24 Gy corresponded with >85% and 95% 1-year LC rates, respectively. For tumors 21 to 30 mm, an 18 Gy single-fraction dose was associated with 75% LC. For tumors 31 to 40 mm, a 15 Gy single-fraction dose yielded ∼69% LC. For 3- to 5-fraction fSRS using doses in the range of 27 to 35 Gy, 80% 1-year LC has been achieved for tumors of 21 to 40 mm in diameter.

Conclusions: TCP for SRS and fSRS are presented. For small lesions ≤20 mm, single doses of ≈18 Gy appear generally associated with excellent rates of LC; for melanoma, higher doses seem warranted. For larger lesions >20 mm, local control rates appear to be ≈ 70% to 75% with usual doses of 15 to 18 Gy, and in this setting, fSRS regimens should be considered. Greater consistency in reporting of dosimetric and LC data is needed to facilitate future pooled analyses. As systemic and biologic therapies evolve, updated analyses will be needed to further assess the necessity, efficacy, and toxicity of SRS and fSRS.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1016/j.ijrobp.2020.10.034DOI Listing
December 2020

Application of radiochromic gel dosimetry to commissioning of a megavoltage research linear accelerator for small-field animal irradiation studies.

Med Phys 2021 Mar 6;48(3):1404-1416. Epub 2021 Feb 6.

Department of Radiation Oncology, University of Michigan, Ann Arbor, MI, USA.

Purpose: To develop and implement an efficient and accurate commissioning procedure for small-field static beam animal irradiation studies on an MV research linear accelerator (Linatron-M9) using radiochromic gel dosimetry.

Materials: The research linear accelerator (Linatron-M9) is a 9 MV linac with a static fixed collimator opening of 5.08 cm diameter. Lead collimators were manually placed to create smaller fields of 2 × 2 cm , 1 × 1 cm , and 0.5 × 0.5 cm . Relative dosimetry measurements were performed, including profiles, percent depth dose (PDD) curves, beam divergence, and relative output factors using various dosimetry tools, including a small volume ionization chamber (A14), GAFCHROMIC™ EBT3 film, and Clearview gel dosimeters. The gel dosimeter was used to provide a 3D volumetric reference of the irradiated fields. The Linatron profiles and relative output factors were extracted at a reference depth of 2 cm with the output factor measured relative to the 2 × 2 cm reference field. Absolute dosimetry was performed using A14 ionization chamber measurements, which were verified using a national standards laboratory remote dosimetry service.

Results: Absolute dosimetry measurements were confirmed within 1.4% (k = 2, 95% confidence = 5%). The relative output factor of the small fields measured with films and gels agreed with a maximum relative percent error difference between the two methods of 1.1 % for the 1 × 1 cm field and 4.3 % for the 0.5 × 0.5 cm field. These relative errors were primarily due to the variability in the collimator positioning. The measured beam profiles demonstrated excellent agreement for beam size (measured as FWHM), within approximately 0.8 mm (or less). Film measurements were more accurate in the penumbra region due to the film's finer resolution compared with the gel dosimeter. Following the van Dyk criteria, the PDD values of the film and gel measurements agree within 11% in the buildup region starting from 0.5 cm depth and within 2.6 % beyond maximum dose and into the fall-off region for depths up to 5 cm. The 2D beam profile isodose lines agree within 0.5 mm in all regions for the 0.5 × 0.5 cm and the 1 × 1 cm fields and within 1 mm for the larger field of 2 × 2 cm . The 2D PDD curves agree within approximately 2% of the maximum in the typical therapy region (1-4 cm) for the 1 × 1 cm and 2 × 2 cm and within 5% for the 0.5 × 0.5 cm field.

Conclusion: This work provides a commissioning process to measure the beam characteristics of a fixed beam MV accelerator with detailed dosimetric evaluation for its implementation in megavoltage small animal irradiation studies. Radiochromic gel dosimeters are efficient small-field relative dosimetry tools providing 3D dose measurements allowing for full representation of dose, dosimeter misalignment corrections and high reproducibility with low inter-dosimeter variability. Overall, radiochromic gels are valuable for fast, full relative dosimetry commissioning in comparison to films for application in high-energy small-field animal irradiation studies.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1002/mp.14685DOI Listing
March 2021

Stereotactic Radiosurgery for Vestibular Schwannomas: Tumor Control Probability Analyses and Recommended Reporting Standards.

Int J Radiat Oncol Biol Phys 2020 Dec 26. Epub 2020 Dec 26.

Machine Learning Department, Moffitt Cancer Center, Tampa, Florida.

Purpose: We sought to investigate the tumor control probability (TCP) of vestibular schwannomas after single-fraction stereotactic radiosurgery (SRS) or hypofractionated SRS over 2 to 5 fractions (fSRS).

Methods And Materials: Studies (PubMed indexed from 1993-2017) were eligible for data extraction if they contained dosimetric details of SRS/fSRS correlated with local tumor control. The rate of tumor control at 5 years (or at 3 years if 5-year data were not available) were collated. Poisson modeling estimated the TCP per equivalent dose in 2 Gy per fraction (EQD2) and in 1, 3, and 5 fractions.

Results: Data were extracted from 35 publications containing a total of 5162 patients. TCP modeling was limited by the absence of analyzable data of <11 Gy in a single-fraction, variability in definition of "tumor control," and by lack of significant increase in TCP for doses >12 Gy. Using linear-quadratic-based dose conversion, the 3- to 5-year TCP was estimated at 95% at an EQD2 of 25 Gy, corresponding to 1-, 3-, and 5-fraction doses of 13.8 Gy, 19.2 Gy, and 21.5 Gy, respectively. Single-fraction doses of 10 Gy, 11 Gy, 12 Gy, and 13 Gy predicted a TCP of 85.0%, 88.4%, 91.2%, and 93.5%, respectively. For fSRS, 18 Gy in 3 fractions (EQD2 of 23.0 Gy) and 25 Gy in 5 fractions (EQD2 of 30.2 Gy) corresponded to TCP of 93.6% and 97.2%. Overall, the quality of dosimetric reporting was poor; recommended reporting guidelines are presented.

Conclusions: With current typical SRS doses of 12 Gy in 1 fraction, 18 Gy in 3 fractions, and 25 Gy in 5 fractions, 3- to 5-year TCP exceeds 91%. To improve pooled data analyses to optimize treatment outcomes for patients with vestibular schwannoma, future reports of SRS should include complete dosimetric details with well-defined tumor control and toxicity endpoints.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1016/j.ijrobp.2020.11.019DOI Listing
December 2020

A Primer on Dose-Response Data Modeling in Radiation Therapy.

Int J Radiat Oncol Biol Phys 2020 Dec 23. Epub 2020 Dec 23.

Department of Machine Learning, Moffitt Cancer Center, Tampa, Florida.

An overview of common approaches used to assess a dose response for radiation therapy-associated endpoints is presented, using lung toxicity data sets analyzed as a part of the High Dose per Fraction, Hypofractionated Treatment Effects in the Clinic effort as an example. Each component presented (eg, data-driven analysis, dose-response analysis, and calculating uncertainties on model prediction) is addressed using established approaches. Specifically, the maximum likelihood method was used to calculate best parameter values of the commonly used logistic model, the profile-likelihood to calculate confidence intervals on model parameters, and the likelihood ratio to determine whether the observed data fit is statistically significant. The bootstrap method was used to calculate confidence intervals for model predictions. Correlated behavior of model parameters and implication for interpreting dose response are discussed.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1016/j.ijrobp.2020.11.020DOI Listing
December 2020

Prostate Stereotactic Body Radiation Therapy: An Overview of Toxicity and Dose Response.

Int J Radiat Oncol Biol Phys 2020 Dec 22. Epub 2020 Dec 22.

Department of Radiation Oncology, University of Kansas, Kansas City, Kansas. Electronic address:

Purpose: Ultrahypofractionationed radiation therapy for prostate cancer is increasingly studied and adopted. The American Association of Physicists in Medicine Working Group on Biological Effects of Hypofractionated Radiotherapy therefore aimed to review studies examining toxicity and quality of life after stereotactic body radiation therapy (SBRT) for prostate cancer and model its effect.

Methods And Materials: We performed a systematic PubMed search of prostate SBRT studies published between 2001 and 2018. Those that analyzed factors associated with late urinary, bowel, or sexual toxicity and/or quality of life were included and reviewed. Normal tissue complication probability modelling was performed on studies that contained detailed dose/volume and outcome data.

Results: We found 13 studies that examined urinary effects, 6 that examined bowel effects, and 4 that examined sexual effects. Most studies included patients with low-intermediate risk prostate cancer treated to 35-40 Gy. Most patients were treated with 5 fractions, with several centers using 4 fractions. Endpoints were heterogeneous and included both physician-scored toxicity and patient-reported quality of life. Most toxicities were mild-moderate (eg, grade 1-2) with a very low overall incidence of severe toxicity (eg, grade 3 or higher, usually <3%). Side effects were associated with both dosimetric and non-dosimetric factors.

Conclusions: Prostate SBRT appears to be overall well tolerated, with determinants of toxicity that include dosimetric factors and patient factors. Suggested dose constraints include bladder V(Rx Dose)Gy <5-10 cc, urethra Dmax <38-42 Gy, and rectum Dmax <35-38 Gy, though current data do not offer firm guidance on tolerance doses. Several areas for future research are suggested.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1016/j.ijrobp.2020.09.054DOI Listing
December 2020

Tumor response prediction in Y radioembolization with PET-based radiomics features and absorbed dose metrics.

EJNMMI Phys 2020 Dec 9;7(1):74. Epub 2020 Dec 9.

Department of Radiology, University of Michigan, Ann Arbor, MI, USA.

Purpose: To evaluate whether lesion radiomics features and absorbed dose metrics extracted from post-therapy Y PET can be integrated to better predict outcomes in microsphere radioembolization of liver malignancies METHODS: Given the noisy nature of Y PET, first, a liver phantom study with repeated acquisitions and varying reconstruction parameters was used to identify a subset of robust radiomics features for the patient analysis. In 36 radioembolization procedures, Y PET/CT was performed within a couple of hours to extract 46 radiomics features and estimate absorbed dose in 105 primary and metastatic liver lesions. Robust radiomics modeling was based on bootstrapped multivariate logistic regression with shrinkage regularization (LASSO) and Cox regression with LASSO. Nested cross-validation and bootstrap resampling were used for optimal parameter/feature selection and for guarding against overfitting risks. Spearman rank correlation was used to analyze feature associations. Area under the receiver-operating characteristics curve (AUC) was used for lesion response (at first follow-up) analysis while Kaplan-Meier plots and c-index were used to assess progression model performance. Models with absorbed dose only, radiomics only, and combined models were developed to predict lesion outcome.

Results: The phantom study identified 15/46 reproducible and robust radiomics features that were subsequently used in the patient models. A lesion response model with zone percentage (ZP) and mean absorbed dose achieved an AUC of 0.729 (95% CI 0.702-0.758), and a progression model with zone size nonuniformity (ZSN) and absorbed dose achieved a c-index of 0.803 (95% CI 0.790-0.815) on nested cross-validation (CV). Although the combined models outperformed the radiomics only and absorbed dose only models, statistical significance was not achieved with the current limited data set to establish expected superiority.

Conclusion: We have developed new lesion-level response and progression models using textural radiomics features, derived from Y PET combined with mean absorbed dose for predicting outcome in radioembolization. These encouraging, but limited results, will need further validation in independent and larger datasets prior to any clinical adoption.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1186/s40658-020-00340-9DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7726084PMC
December 2020

Characterization of the Tumor Immune Microenvironment Identifies M0 Macrophage-Enriched Cluster as a Poor Prognostic Factor in Hepatocellular Carcinoma.

JCO Clin Cancer Inform 2020 10;4:1002-1013

Department of Radiation Oncology, University of Michigan, Ann Arbor, MI.

Purpose: Hepatocellular carcinoma (HCC) is characterized by a poor prognosis and a high recurrence rate. The tumor immune microenvironment in HCC has been characterized as shifted toward immunosuppression. We conducted a genomic data-driven classification of immune microenvironment HCC subtypes. In addition, we demonstrated their prognostic value and suggested a potential therapeutic targeting strategy.

Methods: RNA sequencing data from The Cancer Genome Atlas-Liver Hepatocellular Carcinoma was used (n = 366). Abundance of immune cells was imputed using CIBERSORT and visualized using unsupervised hierarchic clustering. Overall survival (OS) was analyzed using Kaplan-Meier estimates and Cox regression. Differential expression and gene set enrichment analyses were conducted on immune clusters with poor OS and high programmed death-1 (PD-1)/programmed death-ligand 1 (PD-L1) coexpression. A scoring metric combining differentially expressed genes and immune cell content was created, and its prognostic value and immune checkpoint blockade response prediction was evaluated.

Results: Two clusters were characterized by macrophage enrichment, with distinct M0 and M2 subtypes. M2 ( = .038) and M0 ( = .018) were independently prognostic for OS on multivariable analysis. Kaplan-Meier estimates demonstrated that patients in M0 and M2 treated with sorafenib had decreased OS ( = .041), and angiogenesis hallmark genes were enriched in the M0 group. CXCL6 and POSTN were overexpressed in both the M0 and the PD-1/PD-L1 groups. A score consisting of and expression and absolute M0 macrophage content was discriminatory for OS (intermediate: hazard ratio [HR], 1.59; ≤ .001; unfavorable: HR, 2.08; = .04).

Conclusion: Distinct immune cell clusters with macrophage predominance characterize an aggressive HCC phenotype, defined molecularly by angiogenic gene enrichment and clinically by poor prognosis and sorafenib response. This novel immunogenomic signature may aid in stratification of unresectable patients to receive checkpoint inhibitor and antiangiogenic therapy combinations.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1200/CCI.20.00077DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7713549PMC
October 2020

Step-size effect on calculated photon and electron beam Cherenkov-to-dose conversion factors.

Phys Med 2020 Oct 8;78:32-37. Epub 2020 Sep 8.

Medical Physics Unit, McGill University, Montreal, QC H4A 3J1, Canada.

Purpose: Previous work presented and validated in-water Cherenkov emission (CE)-based radiotherapy dosimetry. Condensed history Monte Carlo (MC)-calculated electron beam CE-to-dose conversion with <4π CE detection, however, could exhibit step-size dependence. This work presents a physics update and numerical study of this step-size dependence in photon and electron beams, elucidates the CE generation physics, and guides further research.

Methods: The CE-to-dose conversion, k, is calculated for photons (6X, 15X) and electrons (6E, 20E) on-axis in-water with: θ±δθ∈{90°±90°(4π),90°±5°,45°±45°,90°±45°}, 10 cm equivalent square, 100 cm SSD, 1cm voxel radius and beam-dependent length. Relative deviation from single-scattering (SS) simulation is evaluated on maximum fractional electron step energy loss ESTEPE∈0.01-0.25. Standard uncertainties (k=1, 10histories) are reported. A simplified method considering only the straight step direction is also implemented.

Results: No significant step-size effect (>0.1%) was observed for dose and all k, except for surface dosimetry at 90°±5° (-1.6%±0.5%, 20E), which is not recommended. Electron SS deviation uncertainties (k=1), otherwise, varied from <0.2% overall to <0.1% with large apertures. Photon uncertainties varied from <1.1% overall to <0.2% non-superficially with large apertures. The simplified straight-step method exhibited overall greater deviation from SS, most notably -2.8%±0.1% (6E) and -2.5%±0.4% (20E) superficially with 90°±45°, and -1.4%±0.3% (6X) and -0.6%±0.2% (15X) non-superficially with 90°±5° for ESTEPE∈[0.10,0.25].

Conclusions: We demonstrate step-size independence of newly-implemented correction in EGSnrc directional Cherenkov calculations. This advances clinical CE-based dosimetry and is useful for the general Monte Carlo community.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1016/j.ejmp.2020.08.015DOI Listing
October 2020

Tumor Control Probability Modeling and Systematic Review of the Literature of Stereotactic Body Radiation Therapy for Prostate Cancer.

Int J Radiat Oncol Biol Phys 2020 Sep 6. Epub 2020 Sep 6.

Department of Radiation Oncology, University of Kansas, Kansas City, Kansas.

Purpose: Dose escalation improves localized prostate cancer disease control, and moderately hypofractionated external beam radiation is noninferior to conventional fractionation. The evolving treatment approach of ultrahypofractionation with stereotactic body radiation therapy (SBRT) allows possible further biological dose escalation (biologically equivalent dose [BED]) and shortened treatment time.

Methods And Materials: The American Association of Physicists in Medicine Working Group on Biological Effects of Hypofractionated Radiation Therapy/SBRT included a subgroup to study the prostate tumor control probability (TCP) with SBRT. We performed a systematic review of the available literature and created a dose-response TCP model for the endpoint of freedom from biochemical relapse. Results were stratified by prostate cancer risk group.

Results: Twenty-five published cohorts were identified for inclusion, with a total of 4821 patients (2235 with low-risk, 1894 with intermediate-risk, and 446 with high-risk disease, when reported) treated with a variety of dose/fractionation schemes, permitting dose-response modeling. Five studies had a median follow-up of more than 5 years. Dosing regimens ranged from 32 to 50 Gy in 4 to 5 fractions, with total BED (α/β = 1.5 Gy) between 183.1 and 383.3 Gy. At 5 years, we found that in patients with low-intermediate risk disease, an equivalent doses of 2 Gy per fraction (EQD2) of 71 Gy (31.7 Gy in 5 fractions) achieved a TCP of 90% and an EQD2 of 90 Gy (36.1 Gy in 5 fractions) achieved a TCP of 95%. In patients with high-risk disease, an EQD2 of 97 Gy (37.6 Gy in 5 fractions) can achieve a TCP of 90% and an EQD2 of 102 Gy (38.7 Gy in 5 fractions) can achieve a TCP of 95%.

Conclusions: We found significant variation in the published literature on target delineation, margins used, dose/fractionation, and treatment schedule. Despite this variation, TCP was excellent. Most prescription doses range from 35 to 40 Gy, delivered in 4 to 5 fractions. The literature did not provide detailed dose-volume data, and our dosimetric analysis was constrained to prescription doses. There are many areas in need of continued research as SBRT continues to evolve as a treatment modality for prostate cancer, including the durability of local control with longer follow-up across risk groups, the efficacy and safety of SBRT as a boost to intensity modulated radiation therapy (IMRT), and the impact of incorporating novel imaging techniques into treatment planning.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1016/j.ijrobp.2020.08.014DOI Listing
September 2020

Radiation Fractionation Schedules Published During the COVID-19 Pandemic: A Systematic Review of the Quality of Evidence and Recommendations for Future Development.

Int J Radiat Oncol Biol Phys 2020 Oct 11;108(2):379-389. Epub 2020 Aug 11.

Mercy Hospital, Springfield, Missouri.

Purpose: Numerous publications during the COVID-19 pandemic recommended the use of hypofractionated radiation therapy. This project assessed aggregate changes in the quality of the evidence supporting these schedules to establish a comprehensive evidence base for future reference and highlight aspects for future study.

Methods And Materials: Based on a systematic review of published recommendations related to dose fractionation during the COVID-19 pandemic, 20 expert panelists assigned to 14 disease groups named and graded the highest quality of evidence schedule(s) used routinely for each condition and also graded all COVID-era recommended schedules. The American Society for Radiation Oncology quality of evidence criteria were used to rank the schedules. Process-related statistics and changes in distributions of quality ratings of the highest-rated versus recommended COVID-19 era schedules were described by disease groups and for specific clinical scenarios.

Results: From January to May 2020 there were 54 relevant publications, including 233 recommended COVID-19-adapted dose fractionations. For site-specific curative and site-specific palliative schedules, there was a significant shift from established higher-quality evidence to lower-quality evidence and expert opinions for the recommended schedules (P = .022 and P < .001, respectively). For curative-intent schedules, the distribution of quality scores was essentially reversed (highest levels of evidence "pre-COVID" vs "in-COVID": high quality, 51.4% vs 4.8%; expert opinion, 5.6% vs 49.3%), although there was variation in the magnitude of shifts between disease sites and among specific indications.

Conclusions: A large number of publications recommended hypofractionated radiation therapy schedules across numerous major disease sites during the COVID-19 pandemic, which were supported by a lower quality of evidence than the highest-quality routinely used dose fractionation schedules. This work provides an evidence-based assessment of these potentially practice-changing recommendations and informs individualized decision-making and counseling of patients. These data could also be used to support radiation therapy practices in the event of second waves or surges of the pandemic in new regions of the world.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1016/j.ijrobp.2020.06.054DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7834196PMC
October 2020

Current status of Radiomics for cancer management: Challenges versus opportunities for clinical practice.

J Appl Clin Med Phys 2020 Jul 22;21(7):7-10. Epub 2020 Jul 22.

Department of Radiation Oncology, University of California Davis Cancer Center, Sacramento, CA, USA.

View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1002/acm2.12982DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7386181PMC
July 2020

Tumor Immune Microenvironment Clusters in Localized Prostate Adenocarcinoma: Prognostic Impact of Macrophage Enriched/Plasma Cell Non-Enriched Subtypes.

J Clin Med 2020 Jun 24;9(6). Epub 2020 Jun 24.

Department of Radiation Oncology, University of Michigan, Ann Arbor, MI 48104, USA.

Background: Prostate cancer (PCa) is characterized by significant heterogeneity in its molecular, genomic, and immunologic characteristics.

Methods: Whole transcriptome RNAseq data from The Cancer Genome Atlas of prostate adenocarcinomas ( = 492) was utilized. The immune microenvironment was characterized using the CIBERSORTX tool to identify immune cell type composition. Unsupervised hierarchical clustering was performed based on immune cell type content. Analyses of progression-free survival (PFS), distant metastases, and overall survival (OS) were performed using Kaplan-Meier estimates and Cox regression multivariable analyses.

Results: Four immune clusters were identified, largely defined by plasma cell, CD4 Memory Resting T Cells (CD4 MR), and M0 and M2 macrophage content (CD4 MRPlasma CellM0M2, CD4 MRPlasma CellM0M2, CD4 MRPlasma CellM0M2, and CD4 MRPlasma CellM0M2). The two macrophage-enriched/plasma cell non-enriched clusters (3 and 4) demonstrated worse PFS (HR 2.24, 95% CI 1.46-3.45, = 0.0002) than the clusters 1 and 2. No metastatic events occurred in the plasma cell enriched, non-macrophage-enriched clusters. Comparing clusters 3 vs. 4, in patients treated by surgery alone, cluster 3 had zero progression events ( < 0.0001). However, cluster 3 patients had worse outcomes after post-operative radiotherapy ( = 0.018).

Conclusion: Distinct tumor immune clusters with a macrophage-enriched, plasma cell non-enriched phenotype and reduced plasma cell enrichment independently characterize an aggressive phenotype in localized prostate cancer that may differentially respond to treatment.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.3390/jcm9061973DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7356642PMC
June 2020

An ionizing radiation acoustic imaging (iRAI) technique for real-time dosimetric measurements for FLASH radiotherapy.

Med Phys 2020 Oct 16;47(10):5090-5101. Epub 2020 Aug 16.

Department of Radiation Oncology, University of Michigan, Ann Arbor, MI, 48109, USA.

Purpose: FLASH radiotherapy (FLASH-RT) is a novel irradiation modality with ultra-high dose rates (>40 Gy/s) that have shown tremendous promise for its ability to enhance normal tissue sparing while maintaining comparable tumor cell eradication toconventional radiotherapy (CONV-RT). Due to its extremely high dose rates, clinical translation of FLASH-RT is hampered by risky delivery and current limitations in dosimetric devices, which cannot accurately measure, in real time, dose at deeper tissue. This work aims to investigate ionizing radiation acoustic imaging (iRAI) as a promising image-guidance modality for real-time deep tissue dose measurements during FLASH-RT. The underlying hypothesis is that iRAI can enable mapping of dose deposition with respect to surrounding tissue with a single linear accelerator (linac) pulse precision in real time. In this work, the relationship between iRAI signal response and deposited dose was investigated as well as the feasibility of using a proof-of-concept dual-modality imaging system of ultrasound and iRAI for treatment beam co-localization with respect to underlying anatomy.

Methods: Two experimental setups were used to study the feasibility of iRAI for FLASH-RT using 6 MeV electrons from a modified Varian Clinac. First, experiments were conducted using a single element focused transducer to take a series of point measurements in a gelatin phantom, which was compared with independent dose measurements using GAFchromic film. Secondly, an ultrasound and iRAI dual-modality imaging system utilizing a phased array transducer was used to take coregistered two-dimensional (2D) iRAI signal amplitude images as well as ultrasound B-mode images, to map the dose deposition with respect to surrounding anatomy in an ex vivo rabbit liver model with a single linac pulse precision.

Results: Using a single element transducer, iRAI measurements showed a highly linear relationship between the iRAI signal amplitude and the linac dose per pulse (r  = 0.9998) with a repeatability precision of 1% and a dose resolution error <2.5% in a homogenous phantom when compared to GAFchromic film dose measurements. These phantom results were used to develop a calibration curve between the iRAI signal response and the delivered dose per pulse. Subsequently, a normalized depth dose curve was generated that agreed with film measurements with an RMSE of 0.0243, using correction factors to account for deviations in measurement conditions with respect to calibration. Experiments on the ex-vivo rabbit liver model demonstrated that a 2D iRAI image could be generated successfully from a single linac pulse, which was fused with the B-mode ultrasound image to provide information about the beam position with respect to surrounding anatomy in real time.

Conclusion: This work demonstrates the potential of using iRAI for real-time deep tissue dosimetry in FLASH-RT. Our results show that iRAI signals are linear with dose and can accurately map the delivered radiation dose with respect to soft tissue anatomy. With its ability to measure dose for individual linac pulses at any location within surrounding soft tissue while identifying where that dose is being delivered anatomically in real time, iRAI can be an indispensable tool to enable safe and efficient clinical translation of FLASH-RT.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1002/mp.14358DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7722001PMC
October 2020

National Cancer Institute Workshop on Artificial Intelligence in Radiation Oncology: Training the Next Generation.

Pract Radiat Oncol 2021 Jan-Feb;11(1):74-83. Epub 2020 Jun 13.

Department of Radiation Oncology, University of Michigan, Ann Arbor, Michigan.

Purpose: Artificial intelligence (AI) is about to touch every aspect of radiation therapy, from consultation to treatment planning, quality assurance, therapy delivery, and outcomes modeling. There is an urgent need to train radiation oncologists and medical physicists in data science to help shepherd AI solutions into clinical practice. Poorly trained personnel may do more harm than good when attempting to apply rapidly developing and complex technologies. As the amount of AI research expands in our field, the radiation oncology community needs to discuss how to educate future generations in this area.

Methods And Materials: The National Cancer Institute (NCI) Workshop on AI in Radiation Oncology (Shady Grove, MD, April 4-5, 2019) was the first of 2 data science workshops in radiation oncology hosted by the NCI in 2019. During this workshop, the Training and Education Working Group was formed by volunteers among the invited attendees. Its members represent radiation oncology, medical physics, radiology, computer science, industry, and the NCI.

Results: In this perspective article written by members of the Training and Education Working Group, we provide and discuss action points relevant for future trainees interested in radiation oncology AI: (1) creating AI awareness and responsible conduct; (2) implementing a practical didactic curriculum; (3) creating a publicly available database of training resources; and (4) accelerating learning and funding opportunities.

Conclusion: Together, these action points can facilitate the translation of AI into clinical practice.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1016/j.prro.2020.06.001DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7293478PMC
June 2020

The role of machine and deep learning in modern medical physics.

Med Phys 2020 Jun;47(5):e125-e126

Department of Radiation Oncology, University of North Carolina, Chapel Hill, NC, 277103, USA.

View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1002/mp.14088DOI Listing
June 2020

Introduction to machine and deep learning for medical physicists.

Med Phys 2020 Jun;47(5):e127-e147

Department of Radiation Oncology, University of Michigan, Ann Arbor, MI, 48103, USA.

Recent years have witnessed tremendous growth in the application of machine learning (ML) and deep learning (DL) techniques in medical physics. Embracing the current big data era, medical physicists equipped with these state-of-the-art tools should be able to solve pressing problems in modern radiation oncology. Here, a review of the basic aspects involved in ML/DL model building, including data processing, model training, and validation for medical physics applications is presented and discussed. Machine learning can be categorized based on the underlying task into supervised learning, unsupervised learning, or reinforcement learning; each of these categories has its own input/output dataset characteristics and aims to solve different classes of problems in medical physics ranging from automation of processes to predictive analytics. It is recognized that data size requirements may vary depending on the specific medical physics application and the nature of the algorithms applied. Data processing, which is a crucial step for model stability and precision, should be performed before training the model. Deep learning as a subset of ML is able to learn multilevel representations from raw input data, eliminating the necessity for hand crafted features in classical ML. It can be thought of as an extension of the classical linear models but with multilayer (deep) structures and nonlinear activation functions. The logic of going "deeper" is related to learning complex data structures and its realization has been aided by recent advancements in parallel computing architectures and the development of more robust optimization methods for efficient training of these algorithms. Model validation is an essential part of ML/DL model building. Without it, the model being developed cannot be easily trusted to generalize to unseen data. Whenever applying ML/DL, one should keep in mind, according to Amara's law, that humans may tend to overestimate the ability of a technology in the short term and underestimate its capability in the long term. To establish ML/DL role into standard clinical workflow, models considering balance between accuracy and interpretability should be developed. Machine learning/DL algorithms have potential in numerous radiation oncology applications, including automatizing mundane procedures, improving efficiency and safety of auto-contouring, treatment planning, quality assurance, motion management, and outcome predictions. Medical physicists have been at the frontiers of technology translation into medicine and they ought to be prepared to embrace the inevitable role of ML/DL in the practice of radiation oncology and lead its clinical implementation.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1002/mp.14140DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7331753PMC
June 2020

Machine and deep learning methods for radiomics.

Med Phys 2020 Jun;47(5):e185-e202

Department of Radiation Oncology, University of Michigan, Ann Arbor, MI, 48103, USA.

Radiomics is an emerging area in quantitative image analysis that aims to relate large-scale extracted imaging information to clinical and biological endpoints. The development of quantitative imaging methods along with machine learning has enabled the opportunity to move data science research towards translation for more personalized cancer treatments. Accumulating evidence has indeed demonstrated that noninvasive advanced imaging analytics, that is, radiomics, can reveal key components of tumor phenotype for multiple three-dimensional lesions at multiple time points over and beyond the course of treatment. These developments in the use of CT, PET, US, and MR imaging could augment patient stratification and prognostication buttressing emerging targeted therapeutic approaches. In recent years, deep learning architectures have demonstrated their tremendous potential for image segmentation, reconstruction, recognition, and classification. Many powerful open-source and commercial platforms are currently available to embark in new research areas of radiomics. Quantitative imaging research, however, is complex and key statistical principles should be followed to realize its full potential. The field of radiomics, in particular, requires a renewed focus on optimal study design/reporting practices and standardization of image acquisition, feature calculation, and rigorous statistical analysis for the field to move forward. In this article, the role of machine and deep learning as a major computational vehicle for advanced model building of radiomics-based signatures or classifiers, and diverse clinical applications, working principles, research opportunities, and available computational platforms for radiomics will be reviewed with examples drawn primarily from oncology. We also address issues related to common applications in medical physics, such as standardization, feature extraction, model building, and validation.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1002/mp.13678DOI Listing
June 2020

Oncology Informatics: Status Quo and Outlook.

Oncology 2020 14;98(6):329-331. Epub 2020 May 14.

Department of Radiation Oncology, University of Freiburg - Medical Center, Freiburg, Germany.

Oncology has undergone rapid progress, with emerging developments in areas including cancer stem cells, molecularly targeted therapies, genomic analyses, and individually tailored immunotherapy. These advances have expanded the tools available in the fight against cancer. Some of these have seen broad media coverage resulting in justified public attention. However, these achievements have only been possible due to rapid developments in the expanding field of biomedical informatics and information technology (IT). Artificial intelligence, radiomics, electronic health records, and electronic patient-reported outcome measures (ePROMS) are only a few of the developments enabling further progress in oncology. The promising impact of IT in oncology will only become reality through a multidisciplinary approach to the complex challenges ahead.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1159/000507586DOI Listing
June 2020