Role of surgical margins in transoral robotic surgery: A question yet to be answered

Real-Time Laryngeal Cancer Boundaries Delineation on White Light and Narrow-Band Imaging Laryngoscopy with Deep Learning

OBJECTIVE

To investigate the potential of deep learning for automatically delineating (segmenting) laryngeal cancer superficial extent on endoscopic images and videos.

METHODS

A retrospective study was conducted extracting and annotating white light (WL) and Narrow-Band Imaging (NBI) frames to train a segmentation model (SegMENT-Plus). Two external datasets were used for validation. The model's performances were compared with those of two otolaryngology residents. In addition, the model was tested on real intraoperative laryngoscopy videos.

RESULTS

A total of 3933 images of laryngeal cancer from 557 patients were used. The model achieved the following median values (interquartile range): Dice Similarity Coefficient (DSC) = 0.83 (0.70-0.90), Intersection over Union (IoU) = 0.83 (0.73-0.90), Accuracy = 0.97 (0.95-0.99), Inference Speed = 25.6 (25.1-26.1) frames per second. The external testing cohorts comprised 156 and 200 images. SegMENT-Plus performed similarly on all three datasets for DSC (p = 0.05) and IoU (p = 0.07). No significant differences were noticed when separately analyzing WL and NBI test images on DSC (p = 0.06) and IoU (p = 0.78) and when analyzing the model versus the two residents on DSC (p = 0.06) and IoU (Senior vs. SegMENT-Plus, p = 0.13; Junior vs. SegMENT-Plus, p = 1.00). The model was then tested on real intraoperative laryngoscopy videos.

CONCLUSION

SegMENT-Plus can accurately delineate laryngeal cancer boundaries in endoscopic images, with performances equal to those of two otolaryngology residents. The results on the two external datasets demonstrate excellent generalization capabilities. The computation speed of the model allowed its application on videolaryngoscopies simulating real-time use. Clinical trials are needed to evaluate the role of this technology in surgical practice and resection margin improvement.

LEVEL OF EVIDENCE

III Laryngoscope, 134:2826-2834, 2024.

Oncological and functional outcomes for transoral robotic surgery following previous radiation treatment for upper aerodigestive tract head and neck cancers. A French multicenter GETTEC group study

BACKGROUND

Transoral robotic surgery (TORS) opens new perspectives. We evaluated the outcomes for patients having undergone TORS after previous radiotherapy.

METHODS

A retrospective multicenter study (n = 138) in a previously irradiated area between 2009 and 2020. Survival was assessed with the Kaplan-Meier method. Prognostic factors were evaluated using a chi-squared test, Fisher's test, or Wilcoxon's test.

RESULTS

The median length of hospital stay was 12.5 days. Bleeding was the most frequent postoperative complication (15.2%, n = 22). Prophylactic vessel ligation did not significantly decrease bleeding. Complications were significantly lower for Tis, T1, and N0 tumors. 91.6% (n = 120) of the patients with a perioperative tracheotomy could be decannulated. Larynx was functional for 65.94% of the patients. The median length of follow-up was 26 months. The 5-year overall and relapse-free survival rates were respectively 59.9% and 43.4%.

CONCLUSION

Oncological and functional results confirmed the value of TORS as a treatment in previously irradiated area.

Neo-adjuvant chemotherapy and transoral robotic surgery in locoregionally advanced oropharyngeal cancer

PURPOSE

To analyze the oncological outcomes and patterns of recurrence of patients with locoregionally advanced oropharyngeal squamous cell carcinoma (OPSCC) who underwent neo-adjuvant chemotherapy (NCT) with subsequent transoral robotic surgery (TORS).

METHODS

A single-center retrospective cohort study was performed, including 198 patients (mean age: 58.6, SD: 9.2). The primary outcome was disease-free survival (DFS).

RESULTS

The median follow-up time was 26.5 months (IQR: 16.0-52.0). Estimated DFS rates (95 % CI) at 1 and 3 years were 86.6 % (81.9-91.7), and 81.4 % (75.7-87.6), respectively. Estimated DSS rates (95 % CI) at 1 and 3 years were 96.7 % (94.1-99.3), and 92.6 % (88.4-97.0), respectively. Estimated OS rates (95 % CI) at 1 and 3 years were 96.2 % (93.4-99.0), and 88.7 % (83.4-94.2), respectively. A total of 31 (15.6 %) patients showed a disease relapse after a median time of 8 months (IQR: 4.0-12.0), but only 12 (6 %) patients died of the disease during the study period.

CONCLUSIONS

This study demonstrates that NCT and TORS can obtain excellent tumor control and survival in locoregionally advanced OPSCC. NCT might reduce the need for adjuvant treatments, and randomized clinical trials should be conducted to better define this aspect.

Artificial Intelligence for Upper Aerodigestive Tract Endoscopy and Laryngoscopy: A Guide for Physicians and State-of-the-Art Review

OBJECTIVE

The endoscopic and laryngoscopic examination is paramount for laryngeal, oropharyngeal, nasopharyngeal, nasal, and oral cavity benign lesions and cancer evaluation. Nevertheless, upper aerodigestive tract (UADT) endoscopy is intrinsically operator-dependent and lacks objective quality standards. At present, there has been an increased interest in artificial intelligence (AI) applications in this area to support physicians during the examination, thus enhancing diagnostic performances. The relative novelty of this research field poses a challenge both for the reviewers and readers as clinicians often lack a specific technical background.

DATA SOURCES

Four bibliographic databases were searched: PubMed, EMBASE, Cochrane, and Google Scholar.

REVIEW METHODS

A structured review of the current literature (up to September 2022) was performed. Search terms related to topics of AI, machine learning (ML), and deep learning (DL) in UADT endoscopy and laryngoscopy were identified and queried by 3 independent reviewers. Citations of selected studies were also evaluated to ensure comprehensiveness.

CONCLUSIONS

Forty-one studies were included in the review. AI and computer vision techniques were used to achieve 3 fundamental tasks in this field: classification, detection, and segmentation. All papers were summarized and reviewed.

IMPLICATIONS FOR PRACTICE

This article comprehensively reviews the latest developments in the application of ML and DL in UADT endoscopy and laryngoscopy, as well as their future clinical implications. The technical basis of AI is also explained, providing guidance for nonexpert readers to allow critical appraisal of the evaluation metrics and the most relevant quality requirements.

Single-Port Versus Multiport da Vinci System for Transoral Robotic Surgery of Hypopharyngeal and Laryngeal Carcinoma

OBJECTIVE

Da Vinci single port (SP) has been recently approved for transoral robotic surgery (TORS). Its characteristics make it particularly feasible for laryngeal and hypopharyngeal surgery. We report our experience comparing intra- and postoperative outcomes, technical advantages, and shortcomings of transoral laryngeal and hypopharyngeal resections performed with the da Vinci SP and the da Vinci Si/Xi systems.

STUDY DESIGN

Retrospective database review.

SETTING

Single academic tertiary care hospital.

METHODS

Subjects included adult patients with laryngeal and hypopharyngeal carcinoma who underwent TORS between 2008 and 2022. The SP and multiport (MP) systems were compared in terms of intraoperative times, short-term postoperative outcomes, and TORS-related complications after a propensity score matching.

RESULTS

A total of 185 patients were enrolled (56 SP vs 129 MP patients), and a cohort of 112 patients was analyzed after matching. The docking time was reduced in the SP group (8.84 ± 4.67 vs 6.45 ± 3.11 minutes; p = .003), as well as console time (53.91 ± 29.38 vs 42.70 ± 13.72 minutes; p = .035). Positive margins were more frequent in the MP group (52% vs 43%; p = .34). The mean decannulation time was 1.86 days longer in the SP group (p = .046). No significant differences emerged from the analysis of the duration of hospitalization, enteral feeding, and TORS-related complications.

CONCLUSION

SP safety profile is comparable to that of previous models, while it showed advantages in terms of reduced docking times. Console times were also shortened due to improved maneuverability and field visualization.

Prognostic role of surgical margins in patients undergoing transoral robotic surgery after neo-adjuvant chemotherapy

PURPOSE

To define if positive and close surgical margins are associated to worse prognosis in patients who underwent transoral robotic surgery (TORS) after neoadjuvant chemotherapy (NCT).

METHODS

A retrospective cohort study was carried out at a tertiary referral center. The primary outcome was local-regional control (LRC), and the results were summarized with hazard ratios (HR) and 95% confidence intervals (CIs).

RESULTS

A total of 308 patients (median age: 62.0, IQR: 55.0-68.2) were included. Univariable analysis showed a significant reduced LRC for patients with positive margins (HR = 1.82, 95% CI: 1.02-3.24). However, they were not associated with worse LRC after adjusting for adverse tumor variables (HR = 0.81, 95% CI: 0.40-1.65). ROC analysis was performed on 123 patients with negative margins (AUC: 0.54) measuring an optimal threshold of 1.25 mm (sensitivity = 60.0%; specificity = 50.5%). Univariable analysis showed non-significant differences between close and wide negative margins (HR = 1.44, 95% CI: 0.59-3.54).

CONCLUSIONS

A positive surgical margin is not an independent predictor of tumor control and survival. A threshold of 1.25 mm was identified as the most appropriate to define close margins, but no difference was measured after distinguishing negative margins in close and wide margins.

Neoadjuvant chemotherapy combined with transoral robotic surgery for stage III and IV laryngeal and hypopharyngeal carcinomas

OBJECTIVES

To analyze the oncological and functional outcomes of patients with stage III-IV laryngo-hypopharyngeal cancer who underwent neoadjuvant chemotherapy (NAC) with subsequent transoral robotic surgery (TORS).

MATERIALS AND METHODS

A single-center retrospective cohort study included a total of 100 patients (median age of 67.0) affected by stage III-IV supraglottic or hypopharyngeal cancer. All patients underwent NAC followed by TORS and risk-adjusted adjuvant therapy. The primary outcome was recurrence-free survival (RFS).

RESULTS

The median follow-up time was 24.0 months. Estimated 2-year (95% CI) OS, DSS, and RFS were 75% (66% - 85%), 84% (76% - 92%), and 65% (56% - 76%), respectively. Among the 11 patients who relapsed on the primary site, 3 underwent salvage total laryngectomy, 3 underwent salvage CRT, and the others receive palliation or supportive care. At 6 months from surgery, 17 patients were still tracheostomized or had a stoma retainer, while 15 were gastrostomy dependent. At the Cox multivariable analysis, the clinical stage at presentation, the number of NAC cycles, and the presence of LVI were found to be independently correlated with the RFS.

CONCLUSION

This study demonstrates that NAC followed by TORS can obtain good tumor control, survival, and organ preservation rates in stage III-IV laryngo-hypopharyngeal cancer.

Development of machine learning models for the prediction of positive surgical margins in transoral robotic surgery (TORS)

PURPOSE

To develop machine learning (ML) models for predicting positive margins in patients undergoing transoral robotic surgery (TORS).

METHODS

Data from 453 patients with laryngeal, hypopharyngeal, and oropharyngeal squamous cell carcinoma were retrospectively collected at a tertiary referral center to train (n = 316) and validate (n = 137) six two-class supervised ML models employing 14 variables available pre-operatively.

RESULTS

The accuracy of the six ML models ranged between 0.67 and 0.75, while the measured AUC between 0.68 and 0.75. The ML algorithms showed high specificity (range: 0.75-0.89) and low sensitivity (range: 0.26-0.64) in detecting patients with positive margins after TORS. NPV was higher (range: 0.73-0.83) compared to PPV (range: 0.45-0.63). T classification and tumor site were the most important predictors of positive surgical margins.

CONCLUSIONS

ML algorithms can identify patients with low risk of positive margins and therefore amenable to TORS.