Comparison of different contouring definitions of the rectum as organ at risk (OAR) and dose-volume parameters predicting rectal inflammation in radiotherapy of prostate cancer: which definition to use?

Prior guided deep difference meta-learner for fast adaptation to stylized segmentation

Radiotherapy treatment planning requires segmenting anatomical structures in various styles, influenced by guidelines, protocols, preferences, or dose planning needs. Deep learning-based auto-segmentation models, trained on anatomical definitions, may not match local clinicians' styles at new institutions. Adapting these models can be challenging without sufficient resources. We hypothesize that consistent differences between segmentation styles and anatomical definitions can be learned from initial patients and applied to pre-trained models for more precise segmentation. We propose a Prior-guided deep difference meta-learner (DDL) to learn and adapt these differences. We collected data from 440 patients for model development and 30 for testing. The dataset includes contours of the prostate clinical target volume (CTV), parotid, and rectum. We developed a deep learning framework that segments new images with a matching style using example styles as a prior, without model retraining. The pre-trained segmentation models were adapted to three different clinician styles for post-operative CTV for prostate, parotid gland, and rectum segmentation. We tested the model's ability to learn unseen styles and compared its performance with transfer learning, using varying amounts of prior patient style data (0-10 patients). Performance was quantitatively evaluated using dice similarity coefficient (DSC) and Hausdorff distance. With exposure to only three patients for the model, the average DSC (%) improved from 78.6, 71.9, 63.0, 69.6, 52.2 and 46.3-84.4, 77.8, 73.0, 77.8, 70.5, 68.1, for CTVstyle1, CTVstyle2, CTVstyle3, Parotidsuperficial, Rectumsuperior, and Rectumposterior, respectively. The proposed Prior-guided DDL is a fast and effortless network for adapting a structure to new styles. The improved segmentation accuracy may result in reduced contour editing time, providing a more efficient and streamlined clinical workflow.

Gastrointestinal Toxicity Prediction Not Influenced by Rectal Contour or Dose-Volume Histogram Definition

PURPOSE

Rectal dose delivered during prostate radiation therapy is associated with gastrointestinal toxicity. Treatment plans are commonly optimized using rectal dose-volume constraints, often whole-rectum relative-volumes (%). We investigated whether improved rectal contouring, use of absolute-volumes (cc), or rectal truncation might improve toxicity prediction.

METHODS AND MATERIALS

Patients from the CHHiP trial (receiving 74 Gy/37 fractions [Fr] vs 60 Gy/20 Fr vs 57 Gy/19 Fr) were included if radiation therapy plans were available (2350/3216 patients), plus toxicity data for relevant analyses (2170/3216 patients). Whole solid rectum relative-volumes (%) dose-volume-histogram (DVH), as submitted by treating center (original contour), was assumed standard-of-care. Three investigational rectal DVHs were generated: (1) reviewed contour per CHHiP protocol; (2) original contour absolute volumes (cc); and (3) truncated original contour (2 versions; ±0 and ±2 cm from planning target volume [PTV]). Dose levels of interest (V30, 40, 50, 60, 70, 74 Gy) in 74 Gy arm were converted by equivalent-dose-in-2 Gy-Fr (EQD2α/β= 3 Gy) for 60 Gy/57 Gy arms. Bootstrapped logistic models predicting late toxicities (frequency G1+/G2+, bleeding G1+/G2+, proctitis G1+/G2+, sphincter control G1+, stricture/ulcer G1+) were compared by area-undercurve (AUC) between standard of care and the 3 investigational rectal definitions.

RESULTS

The alternative dose/volume parameters were compared with the original relative-volume (%) DVH of the whole rectal contour, itself fitted as a weak predictor of toxicity (AUC range, 0.57-0.65 across the 8 toxicity measures). There were no significant differences in toxicity prediction for: (1) original versus reviewed rectal contours (AUCs, 0.57-0.66; P = .21-.98); (2) relative- versus absolute-volumes (AUCs, 0.56-0.63; P = .07-.91); and (3) whole-rectum versus truncation at PTV ± 2 cm (AUCs, 0.57-0.65; P = .05-.99) or PTV ± 0 cm (AUCs, 0.57-0.66; P = .27-.98).

CONCLUSIONS

We used whole-rectum relative-volume DVH, submitted by the treating center, as the standard-of-care dosimetric predictor for rectal toxicity. There were no statistically significant differences in prediction performance when using central rectal contour review, with the use of absolute-volume dosimetry, or with rectal truncation relative to PTV. Whole-rectum relative-volumes were not improved upon for toxicity prediction and should remain standard-of-care.

Data drift in medical machine learning: implications and potential remedies

Data drift refers to differences between the data used in training a machine learning (ML) model and that applied to the model in real-world operation. Medical ML systems can be exposed to various forms of data drift, including differences between the data sampled for training and used in clinical operation, differences between medical practices or context of use between training and clinical use, and time-related changes in patient populations, disease patterns, and data acquisition, to name a few. In this article, we first review the terminology used in ML literature related to data drift, define distinct types of drift, and discuss in detail potential causes within the context of medical applications with an emphasis on medical imaging. We then review the recent literature regarding the effects of data drift on medical ML systems, which overwhelmingly show that data drift can be a major cause for performance deterioration. We then discuss methods for monitoring data drift and mitigating its effects with an emphasis on pre- and post-deployment techniques. Some of the potential methods for drift detection and issues around model retraining when drift is detected are included. Based on our review, we find that data drift is a major concern in medical ML deployment and that more research is needed so that ML models can identify drift early, incorporate effective mitigation strategies and resist performance decay.

Autosegmentation based on different-sized training datasets of consistently-curated volumes and impact on rectal contours in prostate cancer radiation therapy

Background and purpose

Autosegmentation techniques are emerging as time-saving means for radiation therapy (RT) contouring, but the understanding of their performance on different datasets is limited. The aim of this study was to determine agreement between rectal volumes by an existing autosegmentation algorithm and manually-delineated rectal volumes in prostate cancer RT. We also investigated contour quality by different-sized training datasets and consistently-curated volumes for retrained versions of this same algorithm.

Materials and methods

Single-institutional data from 624 prostate cancer patients treated to 50–70 Gy were used. Manually-delineated clinical rectal volumes (clinical) and consistently-curated volumes recontoured to one anatomical guideline (reference) were compared to autocontoured volumes by a commercial autosegmentation tool based on deep-learning (v1; n = 891, multiple-institutional data) and retrained versions using subsets of the curated volumes (v32/64/128/256; n = 32/64/128/256). Evaluations included dose-volume histogram metrics, Dice similarity coefficients, and Hausdorff distances; differences between groups were quantified using parametric or non-parametric hypothesis testing.

Results

Volumes by v1-256 (76–78 cm³) were larger than reference (75 cm³) and clinical (76 cm³). Mean doses by v1-256 (24.2–25.2 Gy) were closer to reference (24.2 Gy) than to clinical (23.8 Gy). Maximum doses were similar for all volumes (65.7–66.0 Gy). Dice for v1-256 and reference (0.87–0.89) were higher than for v1-256 and clinical (0.86–0.87) with corresponding Hausdorff comparisons including reference smaller than comparisons including clinical (5–6 mm vs. 7–8 mm).

Conclusion

Using small single-institutional RT datasets with consistently-defined rectal volumes when training autosegmentation algorithms created contours of similar quality as the same algorithm trained on large multi-institutional datasets.

Spatial descriptions of radiotherapy dose: normal tissue complication models and statistical associations

For decades, dose-volume information for segmented anatomy has provided the essential data for correlating radiotherapy dosimetry with treatment-induced complications. Dose-volume information has formed the basis for modelling those associations via normal tissue complication probability (NTCP) models and for driving treatment planning. Limitations to this approach have been identified. Many studies have emerged demonstrating that the incorporation of information describing the spatial nature of the dose distribution, and potentially its correlation with anatomy, can provide more robust associations with toxicity and seed more general NTCP models. Such approaches are culminating in the application of computationally intensive processes such as machine learning and the application of neural networks. The opportunities these approaches have for individualising treatment, predicting toxicity and expanding the solution space for radiation therapy are substantial and have clearly widespread and disruptive potential. Impediments to reaching that potential include issues associated with data collection, model generalisation and validation. This review examines the role of spatial models of complication and summarises relevant published studies. Sources of data for these studies, appropriate statistical methodology frameworks for processing spatial dose information and extracting relevant features are described. Spatial complication modelling is consolidated as a pathway to guiding future developments towards effective, complication-free radiotherapy treatment.

Incidence and dosimetric predictive factors of late rectal toxicity after low-dose-rate brachytherapy combined with volumetric modulated arc therapy in high-risk prostate cancer at a single institution: Retrospective study

PURPOSE

The purpose of this study is to investigate the incidence of rectal toxicity and to identify the associated dosimetric predictive parameters after I-125 seed low-dose-rate brachytherapy (LDR-BT) combined with volumetric modulated arc therapy (VMAT) and dose constraints.

METHODS AND MATERIALS

In total, 110 patients with high-risk prostate cancer received 110 Gy LDR-BT, followed by 45 Gy VMAT. Rectal toxicity was recorded according to Common Terminology Criteria for Adverse Events v.4.03. The dosimetric factors associated with LDR-BT and VMAT were analyzed to determine their relationship with rectal toxicity. Receiver operating characteristic (ROC) curve analysis was performed for ≥ grade 2 (G2) rectal toxicity prediction.

RESULTS

The follow-up duration was 10.1-115.2 months (median 60.5 months). Seven patients had G2 rectal hemorrhage, and none of the patients had grade 3 rectal hemorrhage. In the univariate analysis, the rectal volume receiving 100% of the prescribed dose (rV₁₀₀) (p < 0.001), the dose covering 2 cc of the rectum (rD_2cc) during LDR-BT (p = 0.002), and the combined rD_2cc during LDR-BT and VMAT (p = 0.001) were identified as predictors of G2 rectal hemorrhage. In the ROC curve analysis, the cutoff value was 0.46 cc for rV₁₀₀, 74.0 Gy for rD_2cc, and 86.8 GyEQD₂ for combined rD_2cc.

CONCLUSION

Predictors of late ≥ G2 rectal hemorrhage are rV₁₀₀, rD_2cc, and combined rD_2cc. The incidence of rectal toxicity is low and acceptable in this setting and is highly dependent on the rectal dose of LDR-BT. The use of higher-quality LDR-BT and VMAT dose constraints may further reduce the rate of rectal hemorrhage.