A dual deep neural network for auto-delineation in cervical cancer radiotherapy with clinical validation

Abstract

Abstract Background and purpose: Artificial intelligence (AI) algorithms are capable of automatically detecting contouring boundaries in medical images. However, the algorithms impact on clinical practice of cervical cancer are unclear. We aimed to develop an AI-assisted system for automatic contouring of the clinical target volume (CTV) and organs-at-risk (OARs) in cervical cancer radiotherapy and conduct clinical-based observations. Materials and methods: We first retrospectively collected data of 203 patients with cervical cancer from three groups (A, B, C). The proposed method named as SegNet was developed and trained with different data groups. Quantitative metrics and clinical-based grading were used to evaluate differences between several groups of automatic contours. Then, 20 additional cases were conducted to compare the workload and quality of AI-assisted contours with manual delineation from scratch. Results: For automatic CTVs, SegNet trained with incorporating multi-group data achieved 0.85±0.01, which was statistically better than SegNet independently trained with the single group A (0.82±0.04), B (0.82±0.03) or C (0.81±0.04). The clinical-based grading also showed that SegNet trained with multi-group data obtained better performance of 352/360 relative to it trained with the single group A (334/360), B (333/360) or C (320/360). The manual revision time for automatic CTVs (OARs not yet include) was 9.54±2.42 minutes relative to fully manual delineation with 30.95 ± 15.24 minutes. Conclusion: The proposed SegNet can improve the performance at automatic delineation for cervical cancer radiotherapy by incorporating multi-group data. It is clinically applicable that the AI-assisted system can shorten manual delineation time at no expense of quality.