Clinical evaluation of a deep learning segmentation model including manual adjustments afterwards for locally advanced breast cancer

Introduction

Deep learning (DL) models are increasingly developed for auto-segmentation in radiotherapy. Qualitative analysis is of great importance for clinical implementation, next to quantitative. This study evaluates a DL segmentation model for left- and right-sided locally advanced breast cancer both quantitatively and qualitatively.

Methods

For each side a DL model was trained, including primary breast CTV (CTVp), lymph node levels 1-4, heart, lungs, humeral head, thyroid and esophagus. For evaluation, both automatic segmentation, including correction of contours when needed, and manual delineation was performed and both processes were timed. Quantitative scoring with dice-similarity coefficient (DSC), 95% Hausdorff Distance (95%HD) and surface DSC (sDSC) was used to compare both the automatic (not-corrected) and corrected contours with the manual contours. Qualitative scoring was performed by five radiotherapy technologists and five radiation oncologists using a 3-point Likert scale.

Results

Time reduction was achieved using auto-segmentation in 95% of the cases, including correction. The time reduction (mean ± std) was 42.4% ± 26.5% and 58.5% ± 19.1% for OARs and CTVs, respectively, corresponding to an absolute mean reduction (hh:mm:ss) of 00:08:51 and 00:25:38. Good quantitative results were achieved before correction, e.g. mean DSC for the right-sided CTVp was 0.92 ± 0.06, whereas correction statistically significantly improved this contour by only 0.02 ± 0.05, respectively. In 92% of the cases, auto-contours were scored as clinically acceptable, with or without corrections.

Conclusions

A DL segmentation model was trained and was shown to be a time-efficient way to generate clinically acceptable contours for locally advanced breast cancer.

Clinical evaluation of two AI models for automated breast cancer plan generation

Background

Artificial intelligence (AI) shows great potential to streamline the treatment planning process. However, its clinical adoption is slow due to the limited number of clinical evaluation studies and because often, the translation of the predicted dose distribution to a deliverable plan is lacking. This study evaluates two different, deliverable AI plans in terms of their clinical acceptability based on quantitative parameters and qualitative evaluation by four radiation oncologists.

Methods

For 20 left-sided node-negative breast cancer patients, treated with a prescribed dose of 40.05 Gy, using tangential beam intensity modulated radiotherapy, two model-based treatment plans were evaluated against the corresponding manual plan. The two models used were an in-house developed U-net model and a vendor-developed contextual atlas regression forest model (cARF). Radiation oncologists evaluated the clinical acceptability of each blinded plan and ranked plans according to preference. Furthermore, a comparison with the manual plan was made based on dose volume histogram parameters, clinical evaluation criteria and preparation time.

Results

The U-net model resulted in a higher average and maximum dose to the PTV (median difference 0.37 Gy and 0.47 Gy respectively) and a slightly higher mean heart dose (MHD) (0.01 Gy). The cARF model led to higher average and maximum doses to the PTV (0.30 and 0.39 Gy respectively) and a slightly higher MHD (0.02 Gy) and mean lung dose (MLD, 0.04 Gy). The maximum MHD/MLD difference was ≤ 0.5 Gy for both AI plans. Regardless of these dose differences, 90–95% of the AI plans were considered clinically acceptable versus 90% of the manual plans. Preferences varied between the radiation oncologists. Plan preparation time was comparable between the U-net model and the manual plan (287 s vs 253 s) while the cARF model took longer (471 s). When only considering user interaction, plan generation time was 121 s for the cARF model and 137 s for the U-net model.

Conclusions

Two AI models were used to generate deliverable plans for breast cancer patients, in a time-efficient manner, requiring minimal user interaction. Although the AI plans resulted in slightly higher doses overall, radiation oncologists considered 90–95% of the AI plans clinically acceptable.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13014-022-01993-9.

Elective breast radiotherapy including level I and II lymph nodes: A planning study with the humeral head as planning risk volume

BACKGROUND

The aim of this study was to assess the dose to the humeral head planning risk volume with the currently used high tangential fields (HTF) and compare different planning techniques for breast radiotherapy including axillary level I and II lymph nodes (PTVn) while sparing the humeral head.

METHODS

Ten patients with left-sided breast cancer were enrolled in a planning study with 16 fractions of 2.66 Gy. Four planning techniques were compared: HTF, HTF with sparing of the humeral head, 6-field IMRT with sparing of the humeral head and VMAT with sparing of the humeral head. The humeral head + 10 mm was spared by restricting V40Gy < 1 cc.

RESULTS

The dose to the humeral head was too high with HTF (V40Gy on average 20.7 cc). When sparing the humeral head in HTF, PTVn V90% decreased significantly from 97.9% to 89.4%. 6-field IMRT and VMAT had a PTVn V90% of 98.2% and 99.5% respectively. However, dose to the lungs, heart and especially the contralateral breast increased with VMAT.

CONCLUSIONS

The humeral head is rarely spared when using HTF. When sparing the humeral head, the 6-field IMRT technique leads to adequate PTV coverage while not increasing the dose to the OARs.

Adaptive radiation therapy for breast IMRT-simultaneously integrated boost: three-year clinical experience

PURPOSE

It has been shown that seroma volumes decrease during breast conserving radiotherapy in a significant percentage of patients. We report on our experience with an adaptive radiation therapy (ART) strategy involving rescanning and replanning patients to take this reduction into account during a course of intensity-modulated radiation therapy with simultaneously integrated boost (IMRT-SIB).

MATERIALS

From April 2007 till December 2009, 1274 patients eligible for SIB treatment were enrolled into this protocol. Patients for which the time between the initial planning CT (CT(1)) and lumpectomy was less than 30 days and who had an initial seroma volume >30 cm(3) were rescanned at day 10 of treatment (CT(2)) and replanned when significant changes were observed by the radiation oncologist. Patients received 28 fractions of 1.81 Gy to the breast and 2.30 Gy to the boost volume.

RESULTS

Nine percent (n=113) of the 1274 patients enrolled met the criteria and were rescanned. Of this group, 77% (n=87) of treatment plans were adapted. Time between surgery and CT(1) (20 days versus 20 days for adapted and non-adapted plans, p=0.89) and time between CT(1) and CT(2) (21 days versus 22 days for adapted and non-adapted plans, p=0.43) revealed no procedural differences which might have biased our results. In the adapted plans, seroma decreased significantly from 60 to 27 cm(3) (p<0.001), TBV from 70 to 45 cm(3) (p<0.001) and PTV(boost) from 277 to 220 cm(3) (p<0.001). The volume receiving more than 95% of the boost dose (V(95%(total-dose))) could be reduced by 19% (linear fit, R(2)=0.73) from on average 360 to 292 cm(3) (p<0.001). Delay in treatment and the use of a prolonged treatment schedule with different fractionation for patients with seroma could thus be prevented.

CONCLUSION

The adaptive radiation therapy IMRT-SIB procedure has proven to be efficient and effective, leading to a clinically significant reduction of the high dose volume. Seroma present in a subgroup of patients referred for breast radiation therapy does not hamper the introduction of highly conformal IMRT-SIB techniques.