Alonso, Clayton E.

AbstractPurposeTo reduce the likelihood of errors in organ delineations used for radiotherapy treatment planning, a knowledge‐based quality control (KBQC) system, which discriminates between valid and anomalous delineations is developed.Method and MaterialsThe KBQC is comprised of a group‐wise inference system and anomaly detection modules trained using historical priors from 296 locally advanced lung and prostate cancer patient computational tomographies (CTs). The inference system discriminates different organs based on shape, relational, and intensity features. For a given delineated image set, the inference system solves a combinatorial optimization problem that results in an organ group whose relational features follow those of the training set considering the posterior probabilities obtained from support vector machine (SVM), discriminant subspace ensemble (DSE), and artificial neural network (ANN) classifiers. These classifiers are trained on nonrelational features with a 10‐fold cross‐validation scheme. The anomaly detection module is a bank of ANN autoencoders, each corresponding with an organ, trained on nonrelational features. A heuristic rule detects anomalous organs that exceed predefined organ‐specific tolerances for the feature reconstruction error and the classifier's posterior probabilities. Independent data sets with anomalous delineations were used to test the overall performance of the KBQC system. The anomalous delineations were manually manipulated, computer‐generated, or propagated based on a transformation obtained by imperfect registrations. Both peer‐review–based scoring system and shape similarity coefficient (DSC) were used to label regions of interest (ROIs) as normal or anomalous in two independent test cohorts.ResultsThe accuracy of the classifiers was 99.8%, and the minimum per‐class F1‐scores were 0.99, 0.99, and 0.98 for SVM, DSE, and ANN, respectively. The group‐wise inference system reduced the miss‐classification likelihood for the test data set with anomalous delineations compared to each individual classifier and a fused classifier that used the average posterior probability of all classifiers. For 15 independent locally advanced lung patients, the system detected 79% of the anomalous ROIs. For 1320 auto‐segmented abdominopelvic organs, the anomaly detection system identified anomalous delineations, which also had low Dice similarity coefficient values with respect to manually delineated organs in the training data set.ConclusionThe KBQC system detected anomalous delineations with superior accuracy compared to classification methods that judge only based on posterior probabilities.

625 Background: Leverage the National Cancer Database (NCDB) to evaluate trends in management of squamous cell cancer (SCC) of the rectum and their effect on survival for this uncommon tumor. Methods: Data was obtained from the NCDB for patients diagnosed with SCC of the rectum between 2004 and 2014, including cT1-4, cN0-2, cM0 (cohort A, n = 2,296) tumors. A subgroup analysis was performed on locally advanced tumors (cT1-T2, N+ or cT3, N any, subcohort B, n = 883), treated with chemoradiation (n = 706) or trimodality therapy (n = 177) including chemotherapy, radiation, and surgery. Pathological complete response rate following neoadjuvant therapy was obtained. Univariate and multivariate logistic regression analyses were performed to generate hazard ratios (HR) investigating factors associated with overall survival. Kaplan-Meier (K-M) method was used to estimate overall surviving proportion at 5 and 10 years. Results: The median age was 60 years with a strong female predilection (71% female). Among patients treated with neoadjuvant therapy, 36% achieved a complete pathological response at a median interval of 67 days from completion of radiation therapy to surgery. The K-M estimated 5 and 10 year overall survival for stage I disease was 71.3% and 57.8%, respectively; stage II disease was 57.0% and 38.9%, respectively; stage III disease was 57.8% and 41.5%, respectively. On multivariate analysis, increased age, male gender, more co-morbidities, and higher cT category ( P < 0.001 for each) resulted in worse survival. For locally advanced tumors (subcohort B), there was no difference in survival between chemoradiation alone compared to trimodality therapy ( P = 0.909 on multivariate analysis). Conclusions: Most providers manage locally advanced SCC of the rectum similar to anal cancer, which results in equivalent overall survival and spares patients from the additional morbidity associated with surgical resection. [Table: see text]

AbstractPurposeLeverage the National Cancer Database (NCDB) to evaluate trends in management of nonmetastatic squamous cell cancer (SCC) of the rectum and their effect on survival for this uncommon tumor.Methods and MaterialsRetrospective data was obtained from the NCDB for patients diagnosed with SCC of the rectum between 2004 and 2014, including cT1‐4, cN0‐2, cM0 tumors (cohort A, n = 2296). A subgroup analysis was performed on locally advanced tumors (cT1‐T2, N+ or cT3, N any, subcohort B, n = 883), treated with chemoradiation (n = 706) or trimodality therapy (n = 177) including chemotherapy, radiation, and surgery. Pathological complete response rate following neoadjuvant therapy was obtained. Univariate and multivariate logistic regression analyses were performed to generate hazard ratios (HR) investigating factors associated with overall survival. Kaplan‐Meier (K‐M) method was used to estimate overall surviving proportion at 5 and 10 years.ResultsThe K‐M estimated 5 and 10 year overall survival for stage I disease was 71.3% and 57.8%, respectively; stage II disease was 57.0% and 38.9%, respectively; stage III disease was 57.8% and 41.5%, respectively. On multivariate analysis, higher cT category (P < 0.001) resulted in worse survival. For locally advanced tumors (subcohort B), there was no significant difference in survival between chemoradiation alone compared to trimodality therapy (P = 0.909 on multivariate analysis).ConclusionsMost providers manage locally advanced SCC of the rectum similar to anal cancer, which results in equivalent overall survival and spares patients from the additional morbidity associated with surgical resection.

PurposeTo develop a quality assurance (QA) tool that identifies inaccurate organ at risk (OAR) delineations.MethodsThe QA tool computed volumetric features from prior OAR delineation data from 73 thoracic patients to construct a reference database. All volumetric features of the OAR delineation are computed in three‐dimensional space. Volumetric features of a new OAR are compared with respect to those in the reference database to discern delineation outliers. A multicriteria outlier detection system warns users of specific delineation outliers based on combinations of deviant features. Fifteen independent experimental sets including automatic, propagated, and clinically approved manual delineation sets were used for verification. The verification OARs included manipulations to mimic common errors. Three experts reviewed the experimental sets to identify and classify errors, first without; and then 1 week after with the QA tool.ResultsIn the cohort of manual delineations with manual manipulations, the QA tool detected 94% of the mimicked errors. Overall, it detected 37% of the minor and 85% of the major errors. The QA tool improved reviewer error detection sensitivity from 61% to 68% for minor errors (P = 0.17), and from 78% to 87% for major errors (P = 0.02).ConclusionsThe QA tool assists users to detect potential delineation errors. QA tool integration into clinical procedures may reduce the frequency of inaccurate OAR delineation, and potentially improve safety and quality of radiation treatment planning.

AbstractEarly‐stage nodular lymphocyte‐predominant Hodgkin lymphoma (NLPHL) is associated with a favorable prognosis. Our aim was to evaluate the patterns of care of radiotherapy utilization in this disease and to define the relationship between treatment modality and survival. The National Cancer Database was queried for patients with stages I‐II NLPHL diagnosed from 2004 to 2012. Patients were compared based on primary therapy into four categories: radiotherapy, chemotherapy, both, or neither. Covariate‐adjusted and propensity score‐weighted (PS) Cox proportional hazards models were used, adjusting for potential factors confounding survival. After exclusions, 1420 patients were evaluated, 571 (40%) received radiotherapy alone, 318 (22%) received chemotherapy alone, 351 (25%) received both, and 180 (13%) received neither. Younger patient age (P = 0.001), female gender (P = 0.019), and chemotherapy use (P < 0.001) were associated with decreased radiotherapy utilization. On PS, radiation alone (HR = 0.298, P < 0.001) and chemoradiotherapy (HR = 0.258, P < 0.001) were associated with improved survival compared to no upfront therapy, but the use of chemotherapy alone did not statistically differ compared to no initial therapy (HR = 0.784, P = 0.078). In this large database analysis, over one‐third of patients with early‐stage NLPHL did not receive radiotherapy as a component of initial therapy. The omission of upfront radiotherapy was associated with inferior survival.