Independent test of a model to predict severe acute esophagitis

Abstract

PurposeTreatment planning factors are known to affect the risk of severe acute esophagitis during thoracic radiation therapy. We tested a previously published model to predict the risk of severe acute esophagitis on an independent data set.Methods and materialsThe data set consists of data from patients who had recoverable treatment plans and received definitive radiation therapy for non–small cell carcinoma of the lung at a single institution between November 2004 and January 2010. Complete esophagus dose-volume and available clinical information was extracted using our in-house software. The previously published model was a logistic function with a combination of mean esophageal dose and use of concurrent chemotherapy. In addition to testing the previous model, we used a novel, machine learning-based method to build a maximally predictive model.ResultsNinety-four patients (81.7%) developed Common Terminology Criteria for Adverse Events, Version 4, Grade 2 or more severe esophagitis (Grade 2: n = 79 and Grade 3: n = 15). Univariate analysis revealed that the most statistically significant dose-volume parameters included percentage of esophagus volume receiving ≥40 to 60 Gy, minimum dose to the highest 20% of esophagus volume (D20) to D35, and mean dose. Other significant predictors included concurrent chemotherapy and patient age. The previously published model predicted risk effectively with a Spearman's rank correlation coefficient (rs) of 0.43 (P < .001) with good calibration (Hosmer-Lemeshow goodness of fit: P = .537). A new model that was built from the current data set found the same variables, yielding an rs of 0.43 (P < .001) with a logistic function of 0.0853 × mean esophageal dose [Gy] + 1.49 × concurrent chemotherapy [1/0] − 1.75 and Hosmer-Lemeshow P = .659. A novel preconditioned least absolute shrinkage and selection operator method yielded an average rs of 0.38 on 100 bootstrapped data sets.ConclusionsThe previously published model was validated on an independent data set and determined to be nearly as predictive as the best possible two-parameter logistic model even though it overpredicted risk systematically. A novel, machine learning-based model using a bootstrapping approach showed reasonable predictive power.

Machine learning highlights the deficiency of conventional dosimetric constraints for prevention of high-grade radiation esophagitis in non-small cell lung cancer treated with chemoradiation

“…In our dataset, we observed grade 3 radiation esophagitis in 23 of 202 patients (11.4%), which is lower than historically observed rates of >15% [3,10,35,37]. The comparatively low radiation esophagitis rates we report here may reflect treatment improvements over eras reflecting improved symptom prevention and proactive care, as well as advanced radiation treatment modalities.…”

“…Additionally, the lower than expected toxicity rates may contribute to a lack of reliable predictors in our models due to insufficient radiation esophagitis events. By comparison, other studies reporting robust predictors for grade 2 radiation esophagitis observed toxicity rates upwards of 50% [9,10,35].…”

“…This is in contrast to previously published retrospective [12,32], prospective [16,33] and randomized [34] studies showing associations between esophageal toxicity and predictors including age, tumor nodal stage, concurrent chemotherapy and BMI. It also contrasts with dosimetric factors including eso mean, eso max, as well as eso V20, eso V35, eso V60 which have been analyzed retrospectively in [35], and using a mixture of retrospective and prospective collected datasets in [9]. It is important to note that we specifically examined grade 3 RE, whereas much of the prior literature has examined grade 2 esophagitis, a less clinically impactful toxicity [36,37].…”

“…Previously published reports using models developed using a median cohort size of 141 (well below our actual cohort), and median rate of radiation esophagitis of 11.1%, used logistic regression approaches [1,9,12,17,31,32,34,36,37,39], while some other studies would fit Lyman-Kutcher-Burman (LKB) models, both using pre-IMRT era cohorts [14,33], which may not reflect the current risks of radiation-induced toxicity in a contemporary setting. A more recent study used lasso regularization using a smaller cohort of 94 patients, where 16% developed radiation esophagitis [35]. Finally, some studies with considerably larger cohorts are limited to conformal radiotherapy patients and/or a rather small number of IMRT patients [9,14].…”

Serum Alpha-2-Macroglobulin as an intrinsic radioprotective factor in patients undergoing thoracic radiation therapy

“…Several authors have reported on the increased risk for high-grade esophagitis after sequential and especially concurrent chemo-RT in comparison to RT alone. 5,6,[44][45][46][47] For pneumonitis, we were able to validate the correlation with radiation dose received by the heart (in our model, max dose in heart). Different lung dose volumes (V5-V40 and mean dose in lung) have been found to predict the development of pneumonitis 7,10,48 .…”