Clinical Acceptability of Automatically Generated Elective Lymph Node Volumes for Head and Neck Cancer Patients

PURPOSE/OBJECTIVE(S)

Manual contouring of head and neck lymph node levels is a time-intensive process prone to provider-specific variation. The purpose of this work is to generate a clinical segmentation tool while minimizing the amount of manual effort required by physicians to develop training datasets and review contours. Here we investigate an approach to curate, develop, and clinically validate an auto-contouring model for standard cervical lymph node volumes in the head and neck using a publicly available deep learning architecture. This model updates our previously validated tool to reflect modern practices in lymph node segmentation.

MATERIALS/METHODS

With the assistance of a resident physician, five radiation oncologists manually contoured individual lymph node levels on CT scans for three separate patients treated definitively with radiation or chemoradiation for oropharynx cancer, resulting in 15 unique ground truth cases. These cases were then used to train an nnUnet deep-learning model to generate automated contours for 32 additional cases. These 32 cases were reviewed, manually edited, and used to create the final model. Finally, the model was used to generate contours on the original 15 CT scans (testing cohort), and providers compared these automated contours with the ground-truth (manual) contours. Two blinded studies were performed. In a double-blinded fashion, providers were first asked to select which set of contours they would prefer to use in clinical practice as a starting point for actual cases. Second, they scored each contour on a Likert scale (1-5) to indicate clinical acceptability, ranging from completely unusable to usable without modification.

RESULTS

Across all lymph node levels (IA, IB, II, III, IV, V, RP), average Dice Similarity Coefficient ranged from 0.77 to 0.89 for AI vs manual contours in the testing cohort. These AI and manual lymph node contours were reviewed by 5 physicians each, resulting in 525 preference scores. Across all lymph nodes, the AI contour was superior to or equally preferred to the manual contours at rates ranging from 75% to 91% in the first blinded study. In the second blinded study, physician preference for the manual vs AI contour was statistically different for only the RP contours (p < 0.01). Thus, there was not a significant difference in clinical acceptability for nodal levels I-V for manual versus AI contours. Across all physician-generated contours, 82% were rated as usable with stylistic to no edits, and across all AI-generated contours, 92% were rated as usable with stylistic to no edits.

CONCLUSION

An approach to generate clinically acceptable automated contours for cervical lymph node levels in the head and neck was demonstrated. Furthermore, for nodal levels I-V, there was no significant difference in clinical acceptability in manual vs AI contours. Because we were able to generate and validate a model for each lymph node level individually, the output is applicable to a complete range of disease in which cervical lymph nodes are treated.

Evaluating the Utility of Traditional Bowel Dose Constraints when Treating Abdominal and Pelvic Sarcomas with Preoperative Radiation Therapy

PURPOSE/OBJECTIVE(S)

Abdominopelvic soft tissue sarcomas (AP-STS) are selectively treated with preoperative radiation therapy (RT) followed by surgery. Due to their rarity, most bowel dose constraints are extrapolated from dosimetric studies for other abdominal or pelvic malignancies where concurrent chemotherapy is often given in addition to RT. We sought to investigate associations between dosimetric bowel constraints and the risk for developing acute and long-term toxicities in patients who received preoperative RT for AP-STS.

MATERIALS/METHODS

We performed a retrospective review of patients treated at a tertiary cancer center for non-metastatic AP-STS between 2005 and 2020. Dosimetric parameters for the "bowel bag" organ at risk structure were extracted for each patient. Chi-Square or Fisher's Exact Test, where appropriate, was utilized to compare the proportion of patients who exceeded constraints to the bowel bag contour as per the consensus RT retroperitoneal sarcoma guidelines, and the development of acute and long-term toxicities.

RESULTS

We identified 39 patients with available dosimetric data. The median follow-up was 34 months (IQR 20-47). Approximately half of the tumors were located in the pelvis (n = 20, 51%), and the majority were treated with IMRT/VMAT (n = 35, 90%). 31 patients (80%) presented with de novo disease, and 20% (n = 8) were recurrent presentations but had not received prior RT. The most common histology was leiomyosarcoma (n = 15, 38%), followed by de-differentiated liposarcoma (n = 8, 21%). The median bowel max dose (defined as D0.1cc) was 5309cGy [IQR 5262-5830]. Thirteen patients (33%) exceeded the volumetric V15 Gy <830cm3 bowel bag constraint, and 18 (46%) exceeded the V45 Gy ≤ 195cm3 bowel bag constraint. Overall, 17 patients (44%) had acute grade 1 diarrhea, and 11 (28%) had grade 1-2 nausea. Five patients (13%) had long-term radiation-related toxicities, including 1 patient who developed an enterocutaneous fistula. There was no association between exceeding V15 Gy>830 cm3 (p = 0.31) or V45 Gy≥195cm3 (p = 0.65) bowel bag constraints and developing a long-term RT toxicity. Similarly, exceeding V15 Gy>830 cm3 or V45 Gy≥195cm3 did not lead to increased risks of developing acute diarrhea, nausea, or any other acute RT toxicity (all p>0.05).

CONCLUSION

Traditional volumetric bowel bag dose constraints are frequently exceeded given the large size of AP-STS at presentation and thus the large RT target volumes. Despite a significant proportion of patients exceeding these constraints, these data suggest that RT is overall well-tolerated, and the risk of developing toxicities does not correlate with traditional bowel bag volumetric constraints. There are ongoing efforts to substantially expand this cohort for further investigation into the relationship between dosimetric data and bowel toxicities in order to identify more predictive bowel constraints that can be used during treatment planning of AP-STS.