Search

The purpose of this study is to build and test a support vector machine (SVM) model to predict for the occurrence of lung radiation-induced Grade 2+ pneumonitis. SVM is a sophisticated statistical technique capable of separating the two categories of patients (with/without pneumonitis) using a boundary defined by a complex hypersurface. Despite the complexity, the SVM boundary is only minimally influenced by outliers that are difficult to separate. By contrast, the simple hyperplane boundary computed by the more commonly used and related linear discriminant analysis method is heavily influenced by outliers. Two SVM models were built using data from 219 patients with lung cancer treated using radiotherapy (34 diagnosed with pneumonitis). One model (SVMall) selected input features from all dose and non-dose factors. For comparison, the other model (SVMdose) selected input features only from lung dose-volume factors. Model predictive ability was evaluated using ten-fold cross-validation and receiver operating characteristics (ROC) analysis. For the model SVMall, the area under the cross-validated ROC curve was 0.76 (sensitivity/specificity=74%/75%). Compared to the corresponding SVMdose area of 0.71 (sensitivity/specificity=68%/68%), the predictive ability of SVMall was improved, indicating that non-dose features are important contributors to separating patients with and without pneumonitis. Among the input features selected by model SVMall, the two with highest importance for predicting lung pneumonitis were: (a) generalized equivalent uniform doses close to the mean lung dose, and (b) chemotherapy prior to radiotherapy. The model SVMall is publicly available via internet access. [ABSTRACT FROM AUTHOR]

Traditional methods to compute the tumor control probability (TCP) or normal tissue complication probability (NTCP) typically require a heterogeneous radiation dose distribution to be converted into a simple uniform dose distribution with an equivalent biological effect. Several power-law type dose-volume-histogram reduction schemes, particularly Niemierko’s generalized equivalent uniform dose model [Med. Phys. 26, 1000 (1999)], have been proposed to achieve this goal. In this study, we carefully examine the mathematical outcome of these schemes. We demonstrate that (1) for tumors, with each tumor cell independently responding to local radiation dose, a closed-form analytical solution for tumor survival fraction and TCP can be obtained; (2) for serial structured normal tissues, an exponential power-law form relating survival to functional sub-unit (FSU) radiation is required, and a closed-form analytical solution for the related NTCP is provided; (3) in the case of a parallel structured normal tissue, when NTCP is determined solely by the number of the surviving FSUs, a mathematical solution is available only when there is a non-zero threshold dose and/or a finite critical dose defining the radiotherapy response. Some discussion is offered for the partial irradiation effect on normal tissues in this category; (4) for normal tissues with alternative architectures, where the radiation response of FSU is inhomogeneous, there is no exact global mathematical solution for SF or NTCP within the available schemes. Finally, numerical fits of our models to some experimental data are also presented. [ABSTRACT FROM AUTHOR]

Pulmonary complications, especially idiopathic pneumonitis syndrome (IPS), are potentially life altering or fatal sequelae of hematopoietic cell transplantation (HCT). Total body irradiation (TBI) as part of the conditioning regimen has been implicated in IPS. A comprehensive PENTEC (Pediatric Normal Tissues in the Clinic) review was performed to increase our understanding of the role of TBI in the development of acute, noninfectious IPS. A systematic literature search was conducted using the MEDLINE, PubMed, and Cochrane library databases for articles describing pulmonary toxicity in children treated with HCT. Data pertaining to TBI and pulmonary endpoints were extracted. Risk of IPS was analyzed in relation to patient age, TBI dose, fractionation, dose rate, lung shielding, timing, and type of transplant, with the goal to better understand factors associated with this complication in children undergoing HCT. A logistic regression model was developed using a subset of studies with comparable transplant regimens and sufficient TBI data. Six studies met criteria for modeling of the correlation of TBI parameters with IPS; all consisted of pediatric patients undergoing allogeneic HCT with a cyclophosphamide-based chemotherapy regimen. IPS was variably defined, but all studies that reported IPS were included in this analysis. The mean incidence of post-HCT IPS was 16% (range, 4%-41%). Mortality from IPS, when it occurred, was high (median, 50%; range, 45%-100%). Fractionated TBI prescription doses encompassed a narrow range of 9 to 14 Gy. Many differing TBI methods were reported, and there was an absence of 3-dimensional dose analysis of lung blocking techniques. Thus, a univariate correlation between IPS and total TBI dose, dose fractionation, dose rate, or TBI technique could not be made. However, a model, built from these studies based on prescribed dose using a normalized dose parameter of equivalent dose in 2-Gy fractions (EQD2), adjusted for dose rate, suggested correlation with the development of IPS (P =.0004). The model-predicted odds ratio for IPS was 24.3 Gy–1 (95% confidence interval, 7.0-84.3). Use of TBI lung dose metrics (eg, midlung point dose) could not be successfully modeled, potentially because of dosimetric uncertainties in the actual delivered volumetric lung dose and imperfections in our modeling process. This PENTEC report is a comprehensive review of IPS in pediatric patients receiving fractionated TBI regimens for allogenic HCT. IPS was not clearly associated with 1 single TBI factor. Modeling using dose-rate adjusted EQD2 showed a response with IPS for allogeneic HCT using a cyclophosphamide-based chemotherapy regimen. Therefore, this model suggests IPS mitigation strategies can focus on not just the dose and dose per fraction but also the dose rate used in TBI. More data are needed to confirm this model and to determine the influence of chemotherapy regimens and contribution from graft-versus-host disease. The presence of confounding variables (eg, systemic chemotherapies) that affect risk, the narrow range of fractionated TBI doses found in the literature, and limitations of other reported data (eg, lung point dose) may have prevented a more straightforward link between IPS and total dose from being observed. [ABSTRACT FROM AUTHOR]

Purpose: To present a methodology to use pulmonary gas exchange maps to guide functional avoidance treatment planning in radiation therapy (RT) and evaluate its efficacy compared with ventilation-guided treatment planning.Methods and Materials: Before receiving conventional RT for non-small cell lung cancer, 11 patients underwent hyperpolarized 129Xe gas exchange magnetic resonance imaging to map the distribution of xenon in its gas phase (ventilation) and transiently bound to red blood cells in the alveolar capillaries (gas exchange). Both ventilation and gas exchange maps were independently used to guide development of new functional avoidance treatment plans for every patient, while adhering to institutional dose-volume constraints for normal tissues and target coverage. Furthermore, dose-volume histogram (DVH)-based reoptimizations of the clinical plan, with reductions in mean lung dose (MLD) equal to the functional avoidance plans, were created to serve as the control group. To evaluate each plan (regardless of type), gas exchange maps, representing end-to-end lung function, were used to calculate gas exchange-weighted MLD (fMLD), gas exchange-weighted volume receiving ≥20 Gy (fV20), and mean dose in the highest gas exchanging 33% and 50% volumes of lung (MLD-f33% and MLD-f50%). Using each clinically approved plan as a baseline, the reductions in functional metrics were compared for ventilation-optimization, gas exchange optimization, and DVH-based reoptimization. Statistical significance was determined using the Freidman test, with subsequent subdivision when indicated by P values less than .10 and post hoc testing with Wilcoxon signed rank tests to determine significant differences (P < .05). Toxicity modeling was performed using an established function-based model to estimate clinical significance of the results.Results: Compared with DVH-based reoptimization of the clinically approved plans, gas exchange-guided functional avoidance planning more effectively reduced the gas exchange-weighted metrics fMLD (average ± SD, -78 ± 79 cGy, compared with -45 ± 34 cGy; P = .03), MLD-f33% (-135 ± 136 cGy, compared with -52 ± 47 cGy; P = .004), and MLD-f50% (-96 ± 95 cGy, compared with -47 ± 40 cGy; P = .01). Comparing the 2 functional planning types, Gas Exchange-Guided planning more effectively reduced MLD-f33% compared with ventilation-guided planning (-64 ± 95; P = .009). For some patients, Gas Exchange-Guided functional avoidance plans demonstrated clinically significant reductions in model-predicted toxicity, more so than the accompanying ventilation-guided plans and DVH-based reoptimizations.Conclusion: Gas Exchange-Guided planning effectively reduced dose to high gas exchanging regions of lung while maintaining clinically acceptable plan quality. In many patients, ventilation-guided planning incidentally reduced dose to higher gas exchange regions, to a lesser extent. This methodology enables future prospective trials to examine patient outcomes. [ABSTRACT FROM AUTHOR]

Purpose: Pretreatment quality assurance (QA) of treatment plans often requires a high cognitive workload and considerable time expenditure. This study explores the use of machine learning to classify pretreatment chart check QA for a given radiation plan as difficult or less difficult, thereby alerting the physicists to increase scrutiny on difficult plans., Methods and Materials: Pretreatment QA data were collected for 973 cases between July 2018 and October 2020. The outcome variable, a degree of difficulty, was collected as a subjective rating by physicists who performed the pretreatment chart checks. Potential features were identified based on clinical relevance, contribution to plan complexity, and QA metrics. Five machine learning models were developed: support vector machine, random forest classifier, adaboost classifier, decision tree classifier, and neural network. These were incorporated into a voting classifier, where at least 2 algorithms needed to predict a case as difficult for it to be classified as such. Sensitivity analyses were conducted to evaluate feature importance., Results: The voting classifier achieved an overall accuracy of 77.4% on the test set, with 76.5% accuracy on difficult cases and 78.4% accuracy on less difficult cases. Sensitivity analysis showed features associated with plan complexity (number of fractions, dose per monitor unit, number of planning structures, and number of image sets) and clinical relevance (patient age) were sensitive across at least 3 algorithms., Conclusions: This approach can be used to equitably allocate plans to physicists rather than randomly allocate them, potentially improving pretreatment chart check effectiveness by reducing errors propagating downstream., (© 2023 The Authors.)

Purpose: To assess the impact of prospectively sparing the parotid ducts via MRI sialography on patient reported xerostomia for those receiving definitive radiotherapy (RT) for oropharyngeal squamous cell carcinoma., Methods and Materials: Thirty-eight patients with oropharynx cancer to be treated with definitive RT underwent pre-treatment MRI sialograms to localize their parotid ducts. The parotid ducts were maximally spared during treatment planning. Patients reported symptoms (PRO-CTCAE and QLQ-H&N35) were collected at 6 and 12 months post-RT and compared to a historical cohort who underwent conventional parotid gland mean dose sparing. Regression models were generated using parotid and submandibular gland doses with and without incorporating the dose to the parotid ducts to determine the impact of parotid duct dose on patient reported xerostomia., Results: At 6 months post-RT, 12/26 (46%) patients reported ≥moderate xerostomia when undergoing parotid ductal sparing compared to 43/61 (70%) in the historical cohort (p = 0.03). At 12 months post-RT, 8/22 (36%) patients reported ≥moderate xerostomia when undergoing parotid ductal sparing compared to 34/68(50%) in the historical cohort (p = 0.08). Using nested logistic regression models, the mean parotid duct dose was found to significantly relate to patient reported xerostomia severity at 6 months post-RT (p = 0.04) and trended towards statistical significance at 12 months post-RT (p = 0.09). At both 6 and 12 months post-RT, the addition of mean parotid duct dose significantly improved model fit (p < 0.05)., Conclusions: MRI sialography guided parotid duct sparing appears to reduce the rates of patient-reported xerostomia. Further, logistic regression analysis found parotid duct dose to be significantly associated with patient reported xerostomia. A significant improvement in model fit was observed when adding mean parotid duct dose compared to models that only contain mean parotid gland dose and mean contralateral submandibular gland dose., Competing Interests: Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper., (Copyright © 2022 Elsevier B.V. All rights reserved.)

Chart checking is a time intensive process with high cognitive workload for physicists. Previous studies have partially automated and standardized chart checking, but limited studies implement data-driven approaches to reduce cognitive workload for quality assurance processes. This study aims to evaluate feature selection methods to improve the interpretability and transparency of machine learning models in predicting the degree of difficulty for a pretreatment physics chart check. We compare chi-square, mutual information, feature importance thresholding, and greedy feature selection for four different classifiers. Random forest has the highest performance with SMOTE oversampling using mutual information for feature selection (accuracy 84.0%, AUC 87.0%, precision 80.0%, recall 80.0%). This study demonstrates that feature selection methods can improve model interpretability and transparency.

Purpose: This study aimed to prospectively assess dosimetric and clinical effects of treatment planners having a priori knowledge of the maximum achievable dose sparing for organs at risk (OARs) for patients with oropharynx cancer receiving intensity modulated radiation therapy (RT)., Methods and Materials: We examined patients with oropharynx cancer who were treated in prospective clinical trials between February 2012 and April 2019 at our institution. A tool generating estimates of maximum achievable dose sparing for OARs (feasibility dose-volume histogram [FDVH]) was used clinically starting July 2016. Patients were divided into 2 cohorts: Before (ie, baseline) and after (ie, FDVH-guided) FDVH. Doses received by various OARs were compared with those estimated to be achievable per FDVH, and that difference was defined as the excess of feasible dose (EFD). Patient-reported outcome (PRO) questionnaires were completed at 3, 6, and 12 months after treatment. The baseline and FDVH-guided cohorts were compared in terms of EFD, plan quality metrics, and post-RT PRO assessments., Results: A total of 139 patients were included in the analysis (60 in the baseline cohort, 79 in the FDVH-guided cohort). The FDVH-guided cohort had lower EFD to the contralateral parotid by 4.1 Gy, the ipsilateral parotid by 10.6 Gy, the larynx by 4.3 Gy, the oral cavity by 1.5 Gy, and the contralateral submandibular gland by 0.4 Gy. Plan quality metrics were similar between the cohorts. Less variation of EFD was seen in the FDVH-guided cohort for the parotid glands and contralateral submandibular gland (P < .05). The average post-RT PROs were better in the FVHD cohort versus baseline (particularly at the 6-month timepoint for dry mouth frequency, sticky saliva, meal enjoyment, severity of pain, and Eating Assessment Tool 10 composite [swallowing]; P < .05)., Conclusions: Use of FDVH was associated with improved and less variable OAR sparing for clinically delivered plans. FDVH-guided patients had improved PROs compared with baseline with a variety of outcomes significantly improved at 6 months after treatment., (Copyright © 2021 American Society for Radiation Oncology. Published by Elsevier Inc. All rights reserved.)

Purpose: To determine the relationship between mean oral cavity (OC) dose (treated as a singular organ at risk) to patient reported xerostomia and dysgeusia. In addition, we will examine the relationship between oral cavity substructure doses to patient reported xerostomia and dysgeusia. All patients were treated in the setting of deintensification (60 Gy)., Methods and Materials: In the study, 184 and 177 prospectively enrolled patients for de-escalated chemoradiotherapy (CRT) for human papillomavirus (HPV)-positive oropharyngeal cancer submitted PROs at 6 and 12 months, respectively using Patient-Reported Outcomes version of the Common Terminology Criteria for Adverse Events questionnaire. Patient's OC consisting of the following substructures were segmented: oral tongue, base of tongue, floor of mouth, hard and soft palate, cheek mucosa, and upper and lower lip mucosa. Ordinal logistic regression (no/mild vs moderate vs severe/very severe symptoms) was used to compare organs at risk dosimetry to patient reported xerostomia and dysgeusia at 6 and 12 months. Multivariate ordinal logistic regression models were generated., Results: Mean dose to the contralateral parotid ( P  = .04), OC ( P  = .04), and baseline patient reported xerostomia ( P  = .009) were significantly associated with xerostomia severity at 6 months. Only baseline xerostomia ( P  = .02) and mean dose to the contralateral submandibular gland ( P  = .0001) were significantly associated with xerostomia severity at 12 months. The only significant factor related to dysgeusia at either time point was mean dose to the OC at 12 months ( P  = .009). On examining substructures, the mean dose to the floor of mouth was implicated for the dose relationship to 6-month xerostomia ( P  = .04), and the oral tongue was found to be implicated for the relationship for 12-month dysgeusia ( P  = .04)., Conclusions: The mean dose to the OC was found to relate to xerostomia symptoms at 6 months post-CRT and dysgeusia symptoms at 12 months post-CRT. The mean dose to the floor of mouth and oral tongue appeared to drive this relationship for xerostomia and dysgeusia symptoms, respectively. This work suggests the floor of mouth and oral tongue should be prioritized during planning over the rest of the OC. The effect of OC dose relative to other salivary structures for xerostomia appeared to depend on time post-CRT., (© 2022 The Authors.)

Objective: Hypoxia is a characteristic of many tumors and portends a worse prognosis in lung, cervical, prostate, and rectal cancers. Unlike the others, lung cancers present a unique challenge in measuring hypoxia, with invasive biopsies and higher rates of complications. Noninvasive imaging studies detecting hypoxia using isotopes of copper-diacetyl-bis(N4-methylthiosemicarbazone) ((62)Cu-ATSM) have predicted prognosis and treatment outcomes in some small feasibility trials. These images, however, may not identify all areas of hypoxia. Hence, we hypothesize that the addition of another PET imaging agent, copper-pyruvaldehyde-bis(N4-methylthiosemicarbazone) ((62)Cu-PTSM), which can detect areas of perfusion, can augment the information obtained in (62)Cu-ATSM PET scans., Subjects and Methods: To characterize tumors on the basis of both perfusion and hypoxia, 10 patients were studied using both (62)Cu-ATSM and (62)Cu-PTSM PET scans. In addition, proteomic arrays looking at specific proangiogenic, survival, and proinflammatory targets were assessed., Results: Six of 10 patients had evaluable PET scans. Our initial experience of characterizing lung tumor hypoxia using (62)Cu-ATSM and (62)Cu-PTSM PET scans showed that visualization of areas with hypoxia normalized for perfusion is feasible. All studied tumors exhibited some hypoxia. Despite the small sample size, a positive relationship was noted between epidermal growth factor levels and (62)Cu-ATSM-detected hypoxia., Conclusion: This initial series of (62)Cu-ATSM and (62)Cu-PTSM PET scans shows that evaluating lung masses by visualizing hypoxia and perfusion is a feasible and novel technique to provide more information. Further investigation is warranted to assess the potential role of (62)Cu-ATSM and (62)Cu-PTSM PET techniques combined with proteomics as alternatives to invasive biopsy techniques in clinical care.

Purpose: Abnormalities in single photon emission computed tomography (SPECT) perfusion within the lung and heart are often detected following radiation for tumors in∕around the thorax (e.g., lung cancer or left-sided breast cancer). The presence of SPECT perfusion defects is determined by comparing pre- and post-RT SPECT images. However, RT may increase the density of the soft tissue surrounding the lung∕heart (e.g., chest wall∕breast) that could possibly lead to an "apparent" SPECT perfusion defect due to increased attenuation of emitted photons. Further, increases in tissue effective depth will also increase SPECT photon attenuation and may lead to "apparent" SPECT perfusion defects. The authors herein quantitatively assess the degree of density changes and effective depth in soft tissues following radiation in a series of patients on a prospective clinical study., Methods: Patients receiving thoracic RT were enrolled on a prospective clinical study including pre- and post-RT thoracic computed tomography (CT) scans. Using image registration, changes in tissue density and effective depth within the soft tissues were quantified (as absolute change in average CT Hounsfield units, HU, or tissue thickness, cm). Changes in HU and tissue effective depth were considered as a continuous variable. The potential impact of these tissue changes on SPECT images was estimated using simulation data from a female SPECT thorax phantom with varying tissue densities., Results: Pre- and serial post-RT CT images were quantitatively studied in 23 patients (4 breast cancer, 19 lung cancer). Data were generated from soft tissue regions receiving doses of 20-50 Gy. The average increase in density of the chest was 5 HU (range 46 to -69). The average change in breast density was a decrease of -1 HU (range 13 to -13). There was no apparent dose response in neither the dichotomous nor the continuous analysis. Seventy seven soft tissue contours were created for 19 lung cancer patients. The average change in tissue effective depth was +0.2 cm (range -1.9 to 2.2 cm). The changes in HU represent a <2% average change in tissue density. Based on simulation, the small degree of density and tissue effective depth change is unlikely to yield meaningful changes in either SPECT lung or heart perfusion., Conclusions: RT doses of 20-50 Gy can cause up to a 46 HU increase in soft tissue density 6 months post-RT. Post-RT soft tissue effective depth may increase by 2.0 cm. These modest increases in soft tissue density and effective depth are unlikely to be responsible for the perfusion changes seen on post-RT SPECT lung or heart scans. Further, there was no clear dose response of the soft tissue density changes. Ultimately, the authors findings suggest that prior perfusion reports do reflect changes in the physiology of the lungs and heart.

The aim of this study was to explore similarities between intensity-modulated radiotherapy (IMRT) and intensity-modulated arc therapy (IMAT) techniques in the context of the number of multi-leaf collimator (MLC) segments required to achieve plan objectives, the major factor influencing plan quality. Three clinical cases with increasing complexity were studied: (a) prostate only, (b) prostate and seminal vesicles and (c) prostate and pelvic lymph nodes. Initial 'gold-standard' plans with the maximum possible organ-at-risk sparing were generated for all three cases. For each case, multiple IMRT and IMAT plans were generated with varying intensity levels (IMRT) and arc control points (IMAT), which translate into varying MLC segments in both modalities. The IMAT/IMRT plans were forced to mimic the organ-at-risk sparing and target coverage in the gold-standard plans, thereby only allowing the target dose inhomogeneity to be variable. A higher target dose inhomogeneity (quantified as D5--dose to the highest 5% of target volume) implies that the plan is less capable of modulation. For each case, given a similar number of MLC segments, both IMRT and IMAT plans exhibit similar target dose inhomogeneity, indicating that there is no difference in their ability to provide dose painting. Target dose inhomogeneity remained approximately constant with decreasing segments, but sharply increased below a specific critical number of segments (70, 100, 110 for cases a, b, c, respectively). For the cases studied, IMAT and IMRT plans are similar in their dependence on the number of MLC segments. A minimum critical number of segments are required to ensure adequate plan quality. Future studies are needed to establish the range of minimum critical number of segments for different treatment sites and target-organ geometries.

This work investigates the use of the self-organizing map (SOM) technique for predicting lung radiation pneumonitis (RP) risk. SOM is an effective method for projecting and visualizing high-dimensional data in a low-dimensional space (map). By projecting patients with similar data (dose and non-dose factors) onto the same region of the map, commonalities in their outcomes can be visualized and categorized. Once built, the SOM may be used to predict pneumonitis risk by identifying the region of the map that is most similar to a patient's characteristics. Two SOM models were developed from a database of 219 lung cancer patients treated with radiation therapy (34 clinically diagnosed with Grade 2+ pneumonitis). The models were: SOM(all) built from all dose and non-dose factors and, for comparison, SOM(dose) built from dose factors alone. Both models were tested using ten-fold cross validation and Receiver Operating Characteristics (ROC) analysis. Models SOM(all) and SOM(dose) yielded ten-fold cross-validated ROC areas of 0.73 (sensitivity/specificity = 71%/68%) and 0.67 (sensitivity/specificity = 63%/66%), respectively. The significant difference between the cross-validated ROC areas of these two models (p < 0.05) implies that non-dose features add important information toward predicting RP risk. Among the input features selected by model SOM(all), the two with highest impact for increasing RP risk were: (a) higher mean lung dose and (b) chemotherapy prior to radiation therapy. The SOM model developed here may not be extrapolated to treatment techniques outside that used in our database, such as several-field lung intensity modulated radiation therapy or gated radiation therapy.