TEsoNet: knowledge transfer in surgical phase recognition from laparoscopic sleeve gastrectomy to the laparoscopic part of Ivor-Lewis esophagectomy.

BACKGROUND: Surgical phase recognition using computer vision presents an essential requirement for artificial intelligence-assisted analysis of surgical workflow. Its performance is heavily dependent on large amounts of annotated video data, which remain a limited resource, especially concerning highly specialized procedures. Knowledge transfer from common to more complex procedures can promote data efficiency. Phase recognition models trained on large, readily available datasets may be extrapolated and transferred to smaller datasets of different procedures to improve generalizability. The conditions under which transfer learning is appropriate and feasible remain to be established. METHODS: We defined ten operative phases for the laparoscopic part of Ivor-Lewis Esophagectomy through expert consensus. A dataset of 40 videos was annotated accordingly. The knowledge transfer capability of an established model architecture for phase recognition (CNN + LSTM) was adapted to generate a "Transferal Esophagectomy Network" (TEsoNet) for co-training and transfer learning from laparoscopic Sleeve Gastrectomy to the laparoscopic part of Ivor-Lewis Esophagectomy, exploring different training set compositions and training weights. RESULTS: The explored model architecture is capable of accurate phase detection in complex procedures, such as Esophagectomy, even with low quantities of training data. Knowledge transfer between two upper gastrointestinal procedures is feasible and achieves reasonable accuracy with respect to operative phases with high procedural overlap. CONCLUSION: Robust phase recognition models can achieve reasonable yet phase-specific accuracy through transfer learning and co-training between two related procedures, even when exposed to small amounts of training data of the target procedure. Further exploration is required to determine appropriate data amounts, key characteristics of the training procedure and temporal annotation methods required for successful transferal phase recognition. Transfer learning across different procedures addressing small datasets may increase data efficiency. Finally, to enable the surgical application of AI for intraoperative risk mitigation, coverage of rare, specialized procedures needs to be explored.

Systematic review of learning curves in robot-assisted surgery.

BACKGROUND: Increased uptake of robotic surgery has led to interest in learning curves for robot-assisted procedures. Learning curves, however, are often poorly defined. This systematic review was conducted to identify the available evidence investigating surgeon learning curves in robot-assisted surgery. METHODS: MEDLINE, Embase and the Cochrane Library were searched in February 2018, in accordance with PRISMA guidelines, alongside hand searches of key congresses and existing reviews. Eligible articles were those assessing learning curves associated with robot-assisted surgery in patients. RESULTS: Searches identified 2316 records, of which 68 met the eligibility criteria, reporting on 68 unique studies. Of these, 49 assessed learning curves based on patient data across ten surgical specialties. All 49 were observational, largely single-arm (35 of 49, 71 per cent) and included few surgeons. Learning curves exhibited substantial heterogeneity, varying between procedures, studies and metrics. Standards of reporting were generally poor, with only 17 of 49 (35 per cent) quantifying previous experience. Methods used to assess the learning curve were heterogeneous, often lacking statistical validation and using ambiguous terminology. CONCLUSION: Learning curve estimates were subject to considerable uncertainty. Robust evidence was lacking, owing to limitations in study design, frequent reporting gaps and substantial heterogeneity in the methods used to assess learning curves. The opportunity remains for the establishment of optimal quantitative methods for the assessment of learning curves, to inform surgical training programmes and improve patient outcomes.

ANTECEDENTES: La aceptación creciente de la cirugía robótica ha generado interés en las curvas de aprendizaje para los procedimientos asistidos por robot. Sin embargo, las curvas de aprendizaje a menudo están mal definidas. Esta revisión sistemática se realizó para identificar la evidencia disponible en relación a las curvas de aprendizaje del cirujano en la cirugía asistida por robot. MÉTODOS: En Febrero de 2018, se realizaron búsquedas en MEDLINE, Embase y Cochrane Library, de acuerdo con las recomendaciones PRISMA, junto con búsquedas manuales de congresos clave y de revisiones ya existentes. Los artículos elegibles fueron aquellos que evaluaron las curvas de aprendizaje asociadas con la cirugía asistida por robot efectuada en pacientes. RESULTADOS: Las búsquedas bibliográficas identificaron 2.316 registros de los cuales 68 cumplían los criterios de elegibilidad y correspondían a 68 estudios primarios. De estos 68 estudios, 49 evaluaron las curvas de aprendizaje basadas en datos de pacientes de 10 especialidades quirúrgicas. Los 49 estudios eran todos estudios observacionales, en su mayoría de un solo brazo (35/49 (71%)) e incluían pocos cirujanos. Las curvas de aprendizaje mostraban una notable heterogeneidad, variando entre procedimientos, estudios y parámetros analizados. Los estándares de presentación de informes fueron generalmente deficientes, con solo 17/49 (35%) cuantificando la experiencia previa. Los métodos utilizados para evaluar la curva de aprendizaje fueron heterogéneos, a menudo carecían de validación estadística y usaban terminología ambigua. CONCLUSIÓN: Las estimaciones de la curva de aprendizaje estaban sujetas a una considerable incertidumbre, careciendo de evidencia robusta por las limitaciones en el diseño del estudio, lagunas de información en los artículos y heterogeneidad sustancial en los métodos utilizados para evaluar las curvas de aprendizaje. Queda pendiente establecer métodos cuantitativos óptimos para evaluar las curvas de aprendizaje, informar de los programas de formación quirúrgica y mejorar los resultados del paciente.

Artificial Neural Network Individualised Prediction of Time to Colorectal Cancer Surgery.

AIM: Colorectal cancer pathway targets mandate prompt treatment although practicalities may mean patients wait for surgery. This variable period could be utilised for patient optimisation; however, there is currently no reliable predictive system for time to surgery. If individualised surgical waits were prospectively known, tailored prehabilitation could be introduced. METHODS: A dedicated, prospectively populated elective laparoscopic surgery for colorectal cancer with a curative intent database was utilised. Primary endpoint was the prediction of the individualised waiting time for surgery. A multilayered perceptron artificial neural network (ANN) model was trained and tested alongside uni- and multivariate analyses. RESULTS: 668 consecutive patients were included. 8.5% underwent neoadjuvant chemoradiotherapy. The mean time from diagnosis to surgery was 53 days (95% CI 48.3-57.8). ANN correctly identified those having surgery in <8 (97.7% and 98.8%) and <12 weeks (97.1% and 98.8%) of the training and testing cohorts with area under the receiver operating curves of 0.793 and 0.865, respectively. After neoadjuvant treatment, an ASA physical status score was the most important potentially modifiable risk factor for prolonged waits (normalised importance 64%, OR 4.9, 95% CI 1.5-16). The ANN findings were accurately cross-validated with a logistic regression model. CONCLUSION: Artificial neural networks using demographic and diagnostic data successfully predict individual time to colorectal cancer surgery. This could assist the personalisation of preoperative care including the incorporation of prehabilitation interventions.

Colorectal surgery and enhanced recovery: Impact of a simulation-based care pathway training curriculum.

BACKGROUND: The aim was to determine whether a simulation-based care pathway approach (CPA) curriculum could improve compliance for enhanced recovery programs (ERP), and residents' participation in laparoscopic colorectal surgery (LCS). Indeed, trainee surgeons have limited access to LCS as primary operator, and ERP have improved patients' outcomes in colorectal surgery (CS). METHODS: All residents of our department were trained in a simulation-based CPA: perioperative training consisted in virtual patients built according to guidelines in both ERP and CS, whilst intraoperative training involved a virtual reality simulator curriculum. Twenty consecutive patients undergoing CS were prospectively included before (n=10) and after (n=10) the training. All demographic and perioperative data were prospectively collected, including compliance for ERP. Residents' participation as primary operator in LCS was measured. RESULTS: Five residents (PGY 4-7) were enrolled. None had performed LCS as primary operator. Overall satisfaction and usefulness were both rated 4.5/5, usefulness of pre-, post- and intraoperative training was rated 5/5, 4.5/5 and 4/5, respectively. Residents' participation in LCS significantly improved after the training (0% (0-100) vs. 82.5% (10-100); P=0.006). Pre- and intraoperative data were comparable between groups. Postoperative morbidity was also comparable. Compliance for ERP improved at Day 2 in post-training patients (3 (30%) vs. 8 (80%); P=0.035). Length of stay was not modified. CONCLUSIONS: A simulated CPA curriculum to training in LCS and ERP was correctly implemented. It seemed to improve compliance for ERP, and promoted residents participation as primary operator without adversely altering patients' outcomes.