Cervical (upper) esophageal adenocarcinoma is an extremely rare, aggressive, and biologically distinct subtype of esophageal cancer, with few cases reported in literature. As such, pathophysiology and treatment approaches remain poorly understood. We report a case of a 39-year-old male diagnosed with a locally advanced, unresectable, MMR-deficient cervical esophageal adenocarcinoma driven by double somatic mutation of the MLH1 gene. He was treated with pseudocurative intent treatment through combining standard of care (chemotherapy and immunotherapy) with high-dose radiation therapy. The patient achieved complete remission with no evidence of disease recurrence over three years after initial diagnosis. This case highlights the complex decision making that is involved in the treatment of rare cancers for which data from large randomized controlled trials are not available, and the importance of integrating trial data, tumor pathogenesis, biomarkers, and patient preferences.

OBJECTIVE:This study evaluates the performance of an artificial intelligence predictive clinical decision support system (CheLSEA) in generating chest tube management recommendations. METHODS:From October 2020 to May 2021, 50 adult elective pulmonary resection patients with at least 24 h of chest tube drainage were enrolled in a single-arm, double-anonymized, observational study to evaluate CheLSEA's performance compared with standard chest tube care. Clinical status, digital pleural drainage data, and chest X-ray data were collected prospectively. For each query, CheLSEA generated a recommendation for chest tube removal or maintenance. If maintenance was recommended, CheLSEA generated a removal time prediction. RESULTS:Most patients were female (29 of 47, 62%), smokers (39 of 47, 83%), with a median age of 73 (interquartile range [IQR]: 66 to 77) years, who underwent minimally invasive (44 of 47, 94%) lobectomy (41 of 47, 87%) for primary non-small cell lung cancer (35 of 47, 75%). CheLSEA was queried 174 times, 21% (36 of 174) of which triggered the CheLSEA safeguard system, mostly due to grade 3 or increasing subcutaneous emphysema (20 of 36, 56%). CheLSEA recommended chest tube removal in 9% of remaining requests (12 of 138), 83% of which were safe (10 of 12) and 17% of which were premature by ≤6 h (2 of 12). The remaining 126 queries were answered with chest tube maintenance recommendations up to the optimal removal time (97 of 126, 77%) or shortly thereafter (29 of 126, 23%; median = 17 h, IQR: 17 to 22). When predicting chest tube removal time, 93% of responses (82 of 88) were accurate. CONCLUSIONS:CheLSEA provides safe chest tube management recommendations and can potentially enhance care by reliably emulating expert-level clinical guidance.

Chest X-rays are an inexpensive and widely available imaging modality for diagnosing or monitoring a variety of medical conditions. Given their abundance, healthcare providers could greatly benefit from automated systems capable of screening healthy patients and supporting the diagnosis of pathological cases. Deep learning has become central to such decision-support systems, offering accurate and efficient image classification that can improve clinical workflows and reduce radiologist workload. However, despite the rapid evolution of general-purpose neural architectures, particularly attention-based models, their application to medical imaging remains constrained by limited incorporation of medical domain knowledge. Most existing attention mechanisms optimize only task-specific losses, disregarding crucial anatomical and lesion-location priors, which can hinder generalization and interpretability. In this work, we introduce a fully automated, attention-guided classification framework that integrates medical priors through an on-the-fly segmentation of the lungs, followed by a spatially aware attention loss that directs the network’s focus toward clinically relevant regions. The method requires minimal physician input—only a single annotated X-ray indicating potential lesion areas at initialization and generalizes effectively across patients without relying on absolute bounding-box coordinates. Gradient-based activation mapping is further employed to ensure alignment between attention and lesion-specific regions. Our approach is architecture-agnostic and integrates seamlessly into end-to-end pipelines. Experiments on two medical image datasets demonstrate that the proposed segmentation-enhanced attention loss improves both classification accuracy and representation interpretability compared to the standard cross-entropy loss. The code is available at: https://github.com/rcorizzo/cxr-segmentation-attention/ .

Background: There is ongoing equipoise regarding the use of robotics in the field of thoracic surgery. This study compares the safety and length of stay (LOS) outcomes of robotic-assisted thoracoscopic surgery (RATS) with the current standard of care, video-assisted thoracoscopic surgery (VATS), during the initial adoption phase of RATS at a Canadian tertiary care thoracic surgery center. Methods: We retrospectively reviewed lung and mediastinal resections performed via RATS or VATS from 2022 to 2024. Data were extracted from the institutional electronic medical record. Multiple linear regression was performed to assess the impact of surgical approach on LOS, adjusting for relevant covariates. Adverse event rates were compared using Chi-squared tests. Results: A total of 119 RATS and 390 VATS lung resections were analyzed. Median LOS was 3 days [interquartile range (IQR): 1-5 days] for RATS and 2 days (IQR: 1-5 days) for VATS, with no significant difference. Adverse event rates were 29% (35/119) in the RATS group and 39% (153/390) in the VATS group. For mediastinal resections, 41 RATS and 19 VATS cases were included. Median LOS was significantly shorter for RATS [1 day (IQR: 1-2 days)] compared to VATS [2 days (IQR: 2-5 days)] (P<0.05). Adverse event rates were comparable [29% (12/41) for RATS and 26% (5/19) for VATS]. Conclusions: During the early implementation phase of RATS, outcomes were comparable to the established VATS program with respect to safety and LOS. Notably, RATS was associated with a shorter LOS for mediastinal resections.

Background:Large language models (LLMs) offer a potential solution to the labor-intensive nature of systematic reviews. This study evaluated the ability of the GPT model to identify articles that discuss perioperative risk factors for esophagectomy complications. To test the performance of the model, we tested GPT-4 on narrower inclusion criterion and by assessing its ability to discriminate relevant articles that solely identified preoperative risk factors for esophagectomy. Methods:A literature search was run by a trained librarian to identify studies (n = 1,967) discussing risk factors to esophagectomy complications. The articles underwent title and abstract screening by three independent human reviewers and GPT-4. The Python script used for the analysis made Application Programming Interface (API) calls to GPT-4 with screening criteria in natural language. GPT-4's inclusion and exclusion decision were compared to those decided human reviewers. Results:The agreement between the GPT model and human decision was 85.58% for perioperative factors and 78.75% for preoperative factors. The AUC value was 0.87 and 0.75 for the perioperative and preoperative risk factors query, respectively. In the evaluation of perioperative risk factors, the GPT model demonstrated a high recall for included studies at 89%, a positive predictive value of 74%, and a negative predictive value of 84%, with a low false positive rate of 6% and a macro-F1 score of 0.81. For preoperative risk factors, the model showed a recall of 67% for included studies, a positive predictive value of 65%, and a negative predictive value of 85%, with a false positive rate of 15% and a macro-F1 score of 0.66. The interobserver reliability was substantial, with a kappa score of 0.69 for perioperative factors and 0.61 for preoperative factors. Despite lower accuracy under more stringent criteria, the GPT model proved valuable in streamlining the systematic review workflow. Preliminary evaluation of inclusion and exclusion justification provided by the GPT model were reported to have been useful by study screeners, especially in resolving discrepancies during title and abstract screening. Conclusion:This study demonstrates promising use of LLMs to streamline the workflow of systematic reviews. The integration of LLMs in systematic reviews could lead to significant time and cost savings, however caution must be taken for reviews involving stringent a narrower and exclusion criterion. Future research is needed and should explore integrating LLMs in other steps of the systematic review, such as full text screening or data extraction, and compare different LLMs for their effectiveness in various types of systematic reviews.

Deep learning for medical image classification is extremely important for decision support in medical healthcare settings. General-purpose neural network architectures for image classification have become increasingly sophisticated in recent years. Among them, attention-based models have provided significant advancements in deep learning. However, attention mechanisms adopted thus far focus on minimizing task-specific losses and do not fruitfully exploit medical knowledge, such as lesion-specific characteristics, during the training process, resulting in a potential reduction in accuracy. In this paper, we propose an attention-based approach that leverages knowledge of the localization of specific lesion types to guide the model training process. To this end, gradient-based activation mapping is used for incentivizing models to focus on the right area for a given lesion type. The approach is general since it can be applied to any deep learning architecture in end-to-end model training. Our experiments on two real-world medical image datasets show the ability of our approach to improve the classification performance of popular deep-learning model architectures over the classical cross-entropy loss.

A 31-year-old woman presented with pelvic swelling, dysmenorrhea, dysuria, irregular vaginal bleeding, and catamenial right shoulder pain and was found to have large, complex, bilateral pelvic masses on ultrasound. Magnetic resonance imaging demonstrated severe endometriosis including bilateral ovarian endometriomas, bladder nodules, a vaginal nodule, and multiple liver lesions (Figure 1, axial view, Dixon method, arrow showing endometrioma in the liver). She started an oral contraceptive pill (OCP) and had significant improvement of her symptoms.

ObjectiveTo compare the safety and effectiveness of different surgical approaches in thymectomy: robotics, subxiphoid, lateral video-assisted thoracoscopy surgery (LVATS) and open. MethodologyWe retrospectively reviewed 68 cases of thymectomy with a robot-assisted, subxiphoid, LVATS, open sternotomy or thoracotomy approach for thymic lesions or myasthenia gravis between July 2017 and May 2023 at a single centre. Peri-operative outcomes (operating time, estimated blood loss, conversion rates, R0 resection, adverse events and length of stay [LOS]) were collected. ResultsWe observed six conversions to open (from five LVATS and one robot assisted). The median estimated blood loss was lower for LVATS (100.00 [50.0-100.0] mL) compared with open thymectomies (200.0 [150.0-400.0]; P < .001). No intra-operative adverse events were reported in the robotics, subxiphoid or LVATS groups. In patients with thymic tumours (n = 34), R0 resection was achieved in 100% (2/2) of robotics, 83% of subxiphoid (5/6), 93% (13/14) of LVATS and 75% (n = 9/12) of open cases. The median LOS was shortest for robot assisted (1.0 [interquartile range (IQR) 1.0-3.0]), then subxiphoid (2.0 [IQR 1.0-3.0]), LVATS (2.0 [IQR 1.0-3.0]) then open (5.0 [IQR 4.0-6.0]; P < .001). ConclusionsOur results suggest that with a shorter LOS, robotics, subxiphoid and LVATS thymectomies are safe. Larger size studies are required to compare R0 resection rates between these less invasive surgical approaches.

Objective: There is limited clinical evidence to support any specific parenchymal air leak resolution criteria when using digital pleural drainage devices following lung resection. The aim of this study is to determine an optimal air leak resolution criteria, where duration of chest tube drainage is minimized while avoiding complications from premature chest tube removal. Methods: Airflow data averaged at 10-minute intervals was collected prospectively using a digital pleural drainage device (Thopaz; Medela) in 400 patients from 2015 to 2019. All permutations of air leak resolution criteria from<10 to 100 mL/minute for 4 to 12 hours were applied retrospectively to the pleural drainage data to determine air leak duration, and air leak recurrence frequency and volume. Air leak recurrence indicates potential for rather than occurrence of adverse events. Descriptive statistics were used to identify the optimal criteria based on patient safety (low frequency and volume of air leak recurrences), and efficiency (shortest initial air leak Results: The majority of the 400 patients underwent lobectomy (57% [227 out of 400]), wedge resections (29% [115 out of 400]), or segmentectomies (8% [32 out of 400]) for lung cancer (90% [360 out of 400]). An airflow threshold <50 mL/ minute resulted in longer air leak duration before meeting the criteria for air leak resolution (P < .0001). Air leak recurrence frequency and volume were greater in patients with a monitoring period <8 consecutive hours (P < .0001). Conclusions: When using a digital pleural drainage device, a postoperative air leak resolution criteria<50 mL /minute for 8 consecutive hours was associated with the best safety and efficiency profile. (JTCVS Open 2024;18:360-8)

Background:Postoperative pulmonary complications (PPCs) represent a significant source of morbidity and mortality in surgical patients. Measurement of predicted postoperative forced expiratory volume in the first second (ppo FEV1) may allow for reliable prediction of PPCs and perioperative planning. This study aimed to determine if impaired ppo FEV1 is associated with increased risk of PPCs following oncologic lung resection. Methods:Patients who underwent elective pulmonary resection at The Ottawa Hospital between 2008 and 2018 were evaluated. The presence and severity of PPCs as defined by the Ottawa Thoracic Morbidity & Mortality system were analyzed. The incidence of PPCs was evaluated based on different ppo FEV1 cut-off values (40%, 50%, and 60%), and a multivariable logistic regression was performed to identify predictors of PPCs. Results:Of 1,949 included patients, a thoracoscopic approach (64.4%) was most frequently utilized, and lobectomies represented the most common procedure (60.5%). All cut-off ppo FEV1 values of <40% (P<0.001), <50% (P<0.001), and <60% (P=0.004) were associated with more frequent PPCs (13.0%, 11.6%, and 7.6%, respectively), while only ppo FEV1 <50% showed differences in both minor (P<0.001) and major (P=0.005) PPCs. With ppo FEV1 <50%, differences in PPCs were demonstrated specifically in both thoracoscopic (P=0.03) and open (P=0.003) procedures. On multivariable analysis, ppo FEV1 <50% (P=0.03) and need for operative conversion (P<0.001) independently predicted PPCs. Conclusions:Routine assessment of ppo FEV1 is a practical strategy to identify patients at increased risk of developing PPCs, and can identify candidates for preoperative optimization and postoperative pulmonary support.

Background: Surgical technique plays an essential role in achieving good health outcomes. However, the quality of surgical technique reporting remains heterogeneous. Reporting checklists could help authors to describe the surgical technique more transparently and effectively, as well as to assist reviewers and editors evaluate it more informatively, and promote readers to better understand the technique. We previously developed SUPER (surgical technique reporting checklist and standards) to assist authors in reporting their research that contains surgical technique more transparently. However, further explanation and elaboration of each item are needed for better understanding and reporting practice.Methods: We searched surgical literature in PubMed, Google Scholar and journal websites published up to January 2023 to find multidiscipline examples in various article types for each SUPER item.Results: We explain the 22 items of the SUPER and provide rationales item by item alongside. We provide 69 examples from 53 literature that present optimal reporting of the 22 items. Article types of examples include pure surgical technique, and case reports, observational studies and clinical trials that contain surgical technique. Examples are multidisciplinary, including general surgery, orthopaedical surgery, cardiac surgery, thoracic surgery, gastrointestinal surgery, neurological surgery, oncogenic surgery, and emergency surgery etc.Conclusions: Along with SUPER article, this explanation and elaboration file can promote deeper understanding on the SUPER items. We hope that the article could further guide surgeons and researchers in reporting, and assist editors and peer reviewers in reviewing manuscripts related to surgical technique.

Professionalism is a term that medical students experience frequently throughout their training.In medicine, as in other work settings, standards of professionalism are deeply gendered and racialized.As medical schools have slowly become more diverse, more research is needed on how medical students with minoritized identities experience professionalism in relation to their own identities. To explore this topic, I conducted 49 in-depth interviews with fourth year medical students at allopathic medical schools across the United States. Representation of medical students with historically minoritized racial, gender, and sexual identities was prioritized. Using modified grounded theory methods, I identified four strategies medical students use to manage tension between their social identities and professionalism standards: 1) ruling specialties in or out based on alignment between the professional culture of a specialty and participants’ social identities; 2) constructing a social identity to fit perceived professionalism standards; 3) withholding parts of social identity that conflict with perceived professionalism standards; and 4) resisting conformity to perceived professionalism standards. These findings demonstrate the ways in which today's medical students perceive medicine’s professionalism standards to be gendered, racialized, and heteronormative. Ultimately, these narrow standards place an undue mental and emotional burden on students with minoritized identities to construct social identities that conform to medicine’s traditional standards of professionalism.

Minimally invasive surgical techniques have decreased length of stay (LOS) after pulmonary resections as early as postoperative day 1 reported with no impact in overall patient safety. We have implemented a pilot study exploring the feasibility and safety of postoperative day 0 (same-day) discharge after minimally-invasive lung resection.

Background: This study provides an update to a landmark 2004 report describing demographics, training, and trends in adherence to thoracic surgery practice standards in Canada. Methods: An updated questionnaire was administered to all members of the Canadian Association of Thoracic Surgeons via email (n=142, compared to n=68 in 2004). Our report incorporates internal data from Ontario Health and the Canadian Partnership Against Cancer. Results: Forty-eight surgeons completed the survey (male, 70.8%; mean±standard deviation age, 50.3±9.3 years). This represents a 33.8% response rate, compared to 64.7% in 2004. Most surgeons (69%) served a patient population of over 1 million per center; 32%–34% reported an on-call ratio of 1:4–1:5 days, and the average weekly hours worked was 56.4±11.9. Greater access to dedicated geographic units per center (73% in 2021 vs. 53% in 2004) has improved thoracic-associated services and house staff, notably endoscopy units (100% vs. 91%), with 73% of respondents having access to both endobronchial and endoscopic ultrasound. Access to thoracic radiology has also improved, particularly regarding positron emission tomography scanners per center (76.9% vs. 13%). Annual case volumes for lung (255 vs. 128), esophageal (41 vs. 19), and mediastinal resections (30 vs. 13), along with hiatal hernia repair (45 vs. 20), have increased substantially despite reports of operating room availability and radiology as rate-limiting steps. Conclusion: This survey characterizes compliance with current practice standards, addressing the needs of thoracic surgeons across Canada. Over 85% of respondents were aware of the 2004 compliance paper, and 35% had applied for resources and equipment in response.

Background: Chest tube management aims to balance the risks of early chest tube removal (such as postoperative complications and reinsertion) and detriments of excessive and prolonged drainage (e.g., infection, pain, and increased length of stay). The Chest tube Learning Synthesis and Evaluation Assistant (CheLSEA) is an artificial intelligence-based clinical decision support system, designed to combine, interpret, and learn from postoperative patient monitoring data to provide safe and effective recommendations for healthcare providers managing chest tube care. CheLSEA user-interface is an interactive dashboard developed to access recommendations produced by the system. The purpose of this study was to gain an understanding of healthcare professionals’ perceptions, and patient’s views towards an artificial intelligence-based clinical decision support system for chest tube care. An evaluation to assess the usability of the user-interface was also conducted. Methods: This mixed-methods study was conducted in three phases: (I) a survey of healthcare professionals’ perceptions towards artificial intelligence-based clinical decision support system, (II) usability testing sessions with potential users of the system’s user-interface, using a think-aloud approach followed by interviews with closed and open-ended questions organized in a structured worksheet, and (III) semi-structured interviews with patients to ascertain their views toward the use of artificial intelligence-based clinical decision support system in their chest tube care. Results: Survey results showed an overall positive outlook on the usefulness of CheLSEA in chest tube management and its potential to improve patient care. Healthcare professionals helped identify any challenging elements of CheLSEA’s interface and provided suggestions during usability testing. Interface evaluation interviews generated major themes including visibility, understandability, usability, navigation, workflow, and usefulness. Patient interviews highlighted themes such as optimistic attitudes, implementation considerations, transparent communication with healthcare team, overall trust in the surgeon, and desirable features of artificial intelligence clinical decision support systems (AI-CDSS). Conclusions: For CheLSEA to be functional in a clinical setting, the system must have a user-friendly interface that can be integrated with users’ workflow, meet clinical needs, and undergo continual usability reviews.

Background: Despite the widespread acceptance of safety and oncologic equivalence of minimally invasive thoracic surgery, adoption by thoracic surgeons is lagging. Patient perspectives on minimally invasive thoracic surgery versus open surgical approaches has not been well studied. The aim of this survey was to document patient perspective on pain, complication risks, cosmesis, travel burden, and functional outcomes and their relationship to surgical approach. Methods: From 2012-2017, 201 thoracic surgical patients were prospectively enrolled in this observational cohort study. Participants completed a RAND36 short form health survey and a PPOMITS (patient perspectives on open vs. minimally invasive thoracic surgery) questionnaire. Variables of interest were measured on a continuous visual analog scale. PPOMITS questions were classified into three anatomic regions (neck, chest, and abdomen). Surveys were completed preoperatively, then at 1 and 6 months postoperatively. Chi-squared, Fisher's, and independent t-test were used as appropriate. Results: A total of 201 patients were surveyed. Recovery of indices was similar in both minimally invasive surgery (MIS) and open surgery patients. On average, patients placed greater importance on postoperative pain (6.93; 95% CI: 6.69-7.17) than incision size (4.31; 95% CI: 4.0-4.63, P<0.001) and travel burden (4.35; 95% CI: 4.04-4.66, P<0.001). Risk of complications (7.36; 95% CI: 7.14-7.58) was also given more importance than incision size (P<0.001) and travel burden (P<0.001). Findings were similar at each time point and across body regions. Importance of postoperative pain was similar between both groups regardless of surgical site and timing. RAND SF-36 results indicated a significant decline in physical functioning, role limitations due to physical health, energy level, pain, and social functioning at 1 month. All indices recovered to baseline at 6 months. Conclusions: Early deterioration with recovery of functional outcomes at 6 months were similar regardless of surgical approach. Risk of complications was more important to patients than incision size, pain, and distance traveled for treatment. Our results suggest that patients may be willing to enter randomized trials comparing minimally invasive and open approaches, in regionalized cancer care models.

Background: In this study, we investigate indocyanine green (ICG) dye visualization of the thoracic duct (TD) and conduit perfusion during esophagectomy to reduce anastomotic leak (AL) and chylothorax adverse events (AEs). Methods: Retrospective data of adult patients who underwent esophagectomy for esophageal carcinoma between July 2019 and 2022 were included (n = 105). ICG was delivered intravenously (2 mL, 2.5 mg/mL) to assess conduit perfusion into the small bowel mesentery, inguinal lymph nodes, or foot web spaces for TD visualization using fluorescence imaging. Incidence of TD injury, chylothorax, AL, and AEs were collected. Results: A total of 23 patients received ICG (ICG for TD and perfusion (n = 12) and perfusion only (n = 11)), while 82 patients were controls. TD was visualized in 6 of 12 patients who received ICG for TD. No intraoperative TD injuries or postoperative chylothoraces occurred in these patients. Non-ICG patients had 1 (1.22%) intraoperative TD injury and 10 (12.2%) postoperative chylothoraces (grade I–IIIb). While 10 non-ICG patients (12.2%) developed AL (grade I–IVb), only 2 (8.7%) ICG patients developed AL (grade IIIa). Conclusions: This study demonstrates the utility of ICG fluorescence in intraoperative TD and conduit perfusion assessment for limiting AEs. Standard incorporation of ICG in esophagectomy may help surgeons improve the quality of care in this patient population.