Background With the rapid growth of deep learning research for medical applications comes the need for clinical personnel to be comfortable and familiar with these techniques. Taking a proven approach, we developed a straightforward open-source framework for producing automatic contours for head and neck planning computed tomography studies using a convolutional neural network (CNN). Methods Anonymized studies of 229 patients treated at our clinic for head and neck cancer from 2014 to 2018 were used to train and validate the network. We trained a separate CNN iteration for each of 11 common organs at risk, and then used data from 19 patients previously set aside as test cases for evaluation. We used a commercial atlas-based automatic contouring tool as a comparative benchmark on these test cases to ensure acceptable CNN performance. For the CNN contours and the atlas-based contours, performance was measured using three quantitative metrics and physician reviews using survey and quantifiable correction time for each contour. Results The CNN achieved statistically better scores than the atlas-based workflow on the quantitative metrics for 7 of the 11 organs at risk. In the physician review, the CNN contours were more likely to need minor corrections but less likely to need substantial corrections, and the cumulative correction time required was less than for the atlas-based contours for all but two test cases. Conclusions With this validation, we packaged the code framework and trained CNN parameters and a no-code, browser-based interface to facilitate reproducibility and expansion of the work. All scripts and files are available in a public GitHub repository and are ready for immediate use under the MIT license. Our work introduces a deep learning tool for automatic contouring that is easy for novice personnel to use.
To compare the target volume coverage and normal tissue avoidance for high-dose rate (HDR) brachytherapy boost for cervical cancer between intracavitary (IC) tandem and ovoid (T&O) with additional free-hand interstitial (IS) needles (IC-IS) vs. IC brachytherapy alone under ultrasound guidance. Materials/Methods: We reviewed IC T&O cases with free-hand IS needles for cervical cancer treatment from Nov 2018 to Oct 2019. Ultrasound guidance was used for applicator placement. Treatment planning was CT-based; high-risk clinical target volumes (HR-CTVs) and organs at risk (OARs) were delineated before each fraction. Minimum acceptable target coverage was 90% of the prescription dose delivered to the HR-CTV (D90). If these objectives were unmet with the standard IC applicator, IS needles were added at the discretion of the physician. For this analysis, each IC-IS case was re-planned (r) with IC alone and again with IC-IS, matching the dose-limiting normal tissue structure for that treatment. HR-CTV with D90 along with OARs D2cc means and 95% confidence intervals were evaluated between rIC-IS and rIC plans using a paired t-tests, p Z 0.05. Acute side effects were evaluated using CTCAE v5.0 criteria. Results: We identified 6 patients with 20 IC-IS treatments after whole pelvis radiation. Between two to six IS needles were placed for each treatment. All fractions were planned for 7 Gy. Mean D90 per fraction to the HR-CTV for iIC-IS, rIC, and rIC-IS needles was 7.6 Gy (6.3e9.4) vs. 7.1 Gy (4.9e9.3) vs. 7.8 Gy (6.3e9.8) (p Z 0.0000008), respectively, with a relative percent-dose increase of 10.6% from rIC alone to rIC-IS. Regarding OARs, mean D2cc per fraction for iIC-IS, rIC alone, and rIC-IS were: Bladder-5.3 Gy (3.7e6.4) vs. 5.5 (4.1e6.2) vs. 5.5 (3.93 e 6.19) (p Z 0.78); Rectum-3.1 Gy (2.2e4.2) vs. 3.2 (2.1e4.1) vs. 3.3 Gy (2.2e4.1) (p Z 0.14); Sigmoid-3.4 Gy (2.3e4.2) vs. 3.4 (2.5e4.2) vs. 3.4 (2.5e 4.2) (p Z 0.49). Dosimetric results demonstrated in the table. All patients tolerated the procedure well without any CTCAE v5.0 Grade 3 to 5 adverse events. Conclusion: For appropriately selected patients, adding IS needless to IC HDR brachytherapy boost can improve dose coverage of HR-CTV for cervical cancer while not increasing dose to OARs. Long term follow-up is warranted for evaluating differences in clinical outcomes and associated toxicity.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.