Background and purpose

Chest wall toxicity is observed after stereotactic body radiation therapy (SBRT) for peripherally located lung tumors. We utilize machine learning algorithms to identify toxicity predictors to develop dose-volume constraints.

Materials and methods

Twenty-five patient, tumor, and dosimetric features were recorded for 197 consecutive patients with Stage I NSCLC treated with SBRT, 11 of whom (5.6%) developed CTCAEv4 grade ≥2 chest wall pain. Decision tree modeling was used to determine chest wall syndrome (CWS) thresholds for individual features. Significant features were determined using independent multivariate methods. These methods incorporate out-of-bag estimation using Random forests (RF) and bootstrapping (100 iterations) using decision trees.

Results

Univariate analysis identified rib dose to 1 cc < 4000 cGy (P = 0.01), chest wall dose to 30 cc < 1900 cGy (P = 0.035), rib Dmax < 5100 cGy (P = 0.05) and lung dose to 1000 cc < 70 cGy (P = 0.039) to be statistically significant thresholds for avoiding CWS. Subsequent multivariate analysis confirmed the importance of rib dose to 1 cc, chest wall dose to 30 cc, and rib Dmax. Using learning-curve experiments, the dataset proved to be self-consistent and provides a realistic model for CWS analysis.

Conclusions

Using machine learning algorithms in this first of its kind study, we identify robust features and cutoffs predictive for the rare clinical event of CWS. Additional data in planned subsequent multicenter studies will help increase the accuracy of multivariate analysis.

BACKGROUND: The optimal upfront treatment modality for patients with nonmetastatic Gleason Score 9 and 10 prostate cancer (GS 9-10 PCa) is unknown. METHODS: We conducted a retrospective cohort study of patients in the Veterans Health Administration (VHA) with GS 9-10 PCa treated with radical prostatectomy (RP) or external beam radiation therapy with androgen deprivation therapy (EBRT+ADT) from 1/2000 to 12/2010. Outcomes included overall survival (OS), distant metastasis-free survival (DMFS), and salvage/adjuvant therapy-free survival (SAFS), as assessed by Kaplan-Meier analysis. RESULTS: We identified 1220 veterans with GS 9-10 PCa; 335 were treated with RP, and 885 were treated with EBRT+ADT. With a median follow-up of 9.9 years, propensity score-matched analyses demonstrated that RP had superior 10-year OS (70.8% [RP] vs. 61.2% [EBRT+ADT], p < 0.001), 10-year DMFS rates were similar between RP (76.7%) and EBRT+ADT (81.0%), and 10-year SAFS rates were lower for RP vs EBRT + ADT (35.2% [RP] vs. 75.2% [EBRT+ADT], p < 0.001). The receipt of salvage ADT was higher with upfront RP (51.9% vs. 26.1%, p < 0.001), despite receipt of adjuvant/salvage EBRT in 41.8% of RP patients. Among patients treated with RP, there were no differences in outcomes by race. However, higher survival rates were noted among Black patients treated with EBRT+ADT compared with White patients. CONCLUSIONS: This analysis demonstrated higher 10-year OS rates among men treated with upfront RP versus EBRT+ADT, though missing confounders and similar DMFS rates suggest the long-term cause-specific OS rates may be similar. We also highlight real-world outcomes of a diverse patient population in the VHA and improved outcomes for Black patients receiving EBRT+ADT.