The impact of LIMBUS AI based contouring on the efficiency of prostate radiotherapy planning
Kishen Patel,
United Kingdom
PO-1659
Abstract
The impact of LIMBUS AI based contouring on the efficiency of prostate radiotherapy planning
Authors: Kishen Patel1,2, Pandora Rudd3, Kieran Palmer1, Alison Starke4, Jackie Poxon4, Paula Wells5, Karen Tipples5, Niall MacDougall4
1St Bartholomew's Hospital , Clinical Oncology , London, United Kingdom; 2University College London, Cancer Institute, Faculty of Medical Sciences, London, United Kingdom; 3St Bartholomew's Hospital , Clinical Oncology, London, United Kingdom; 4St Bartholomew's Hospital, Physics, London, United Kingdom; 5St Bartholomew's Hospital, Clinical Oncology, London, United Kingdom
Show Affiliations
Hide Affiliations
Purpose or Objective
One in eight men in the UK will be diagnosed with prostate cancer and of those, 30% will have radiotherapy as part of their treatment.
Due to the volume of patients being treated, efficient workflow is vital to the prostate radiotherapy treatment pathway. Expert manual contouring (EC) of the prostate gland, draining lymph nodes and organs at risk (OAR), is required; which is a time consuming process.
Deep learning based automatic contouring artificial intelligence (AI) software, such as Limbus AI, has been developed to help alleviate this workload pressure by automatically generating contours required for the radiotherapy planning process. We aimed to assess whether AI generated contours (AIC) can be used to reduce prostate radiotherapy contouring times compared with EC and the technical quality of the resulting AI generated contours.
Material and Methods
Two Barts Health clinical oncology consultants recorded their contouring times taken for the prostate and seminal vesicles (CTV), lymph nodes (LN) and each OAR for 10 prostate (PR) and 10 prostate and lymph node plans (PLN). Limbus AI version 1.0.6 was used to generate 10 PR and 10 PLN plans from previously treated patients. Time taken by each consultant to amend each AIC to optimal volume was recorded.
The secondary outcome was to assess the geometric similarity between an EC and AIC for each patient. Contours were compared using the dice similarity coefficient (DSC). DSC represents the overlap of volumes with a value of 0 indicating no overlap and 1 a perfect overlap. Values >0.7 suggest excellent agreement.
Results
The median time to complete all EC for a prostate plan was 26 minutes. The median time to amend an AIC PR contour set was 7 minutes, representing a time saving benefit of 19 minutes: a 73% reduction in contouring time.
The median time to complete all EC for prostate and nodes was 1 hour 4 minutes. The median time to amend an AIC PLN contour set was 12 minutes, representing a time saving benefit of 52 minutes: an 81% reduction in contouring time.
Table 1. EC times compared to amending AIC times
| Bladder | Femur L | Femur R | Pelvis LN | CTV | Rectum | Bowel |
| Median Time EC (mins) | 4.4 | 2.1 | 2.1 | 20.0 | 11.6 | 5.4 | 12.0 |
| Median Time AIC (mins) | 0.5 | 0.8 | 0.7 | 5.0 | 4.1 | 1.0 | 3.0
|
The median DSC for all OARs was > 0.8. CTV scored 0.88 and LN, 0.71.
Conclusion
Our results have shown that editing a Limbus auto-contour to a clinical standard was more time efficient than manually contouring structures for prostate plans, with a significant time saving benefit. DSC values showed good agreement between AIC and EC from the Limbus system. For all structures, substantial time was saved editing an AIC, when compared to generating a new EC. Our study shows that AI can safely be used as a substantial time saver in the prostate radiotherapy planning process. Larger studies are required to confirm these preliminary results.