Robotic Extended Right Hepatectomy for Colorectal Liver Metastasis.

BACKGROUND: Advantages of minimally invasive compared with open hepatobiliary surgery include quicker functional recovery, decreased postoperative length of stay, and decreased postoperative opioid use. As more complex operations are approached in minimally invasive fashion, it is imperative to maintain safety and excellent oncologic outcomes. METHODS: In this video, we demonstrate the key principles in performing a safe robotic extended right hepatectomy for colorectal liver metastasis following sound oncologic principles. RESULTS: Key preoperative considerations include (1) early referral to a hepatobiliary surgeon, (2) careful review of cross-sectional imaging to identify the relationship of tumors to major vasculature and any aberrant vascular anatomy, and (3) liver volumetry for every right hepatectomy to determine the need for future liver remnant volume augmentation. Key intraoperative techniques include (1) liberal use of ultrasound before and during transection to determine the relationship of major vasculature to tumor to preserve liver parenchyma without compromising tumor margins, (2) external retraction with vessel loops placed on either side of the transection line as stay sutures to facilitate parenchymal transection, and (3) a crush clamp technique to safely identify and control crossing vessels while dividing liver parenchyma. CONCLUSIONS: With proper preoperative planning and intraoperative use of these techniques, the benefits of a minimally invasive approach can be achieved while maintaining excellence in surgical quality and safety.

Iterative Changes in Risk-Stratified Pancreatectomy Clinical Pathways and Accelerated Discharge After Pancreaticoduodenectomy.

BACKGROUND: Previous implementation of risk-stratified pancreatectomy clinical pathways (RSPCPs) decreased length of stay (LOS) following pancreaticoduodenectomy (PD). This study's primary aim was to measure the association of iterative RSPCP revisions with accelerated discharge and early postoperative outcomes. METHODS: This is a retrospective cohort study of a prospectively maintained surgical database (10/2016-9/2020). In February 2019, revised RSPCPs were implemented with earlier nasogastric tube (NGT) removal (postoperative day [POD] 1 for low risk; POD 2 for high risk) and updated drain fluid amylase cutoffs for POD 1/POD 3 removal. Perioperative outcomes between original and revised pathways were compared. Predictors of accelerated discharge (defined as ≤ POD 5 for low risk; ≤ POD 6 for high risk) were identified. RESULTS: There were 233 (36% high risk) patients in original and 131 (32% high risk) in revised RSPCPs. After revision, the rate of POD 1 NGT removal was higher while POD ≤ 3 drain removal was similar. Median LOS decreased for low risk (5 vs. 6 days, p = 0.011) and high risk (6 vs. 9 days, p = 0.005) with no increase in delayed gastric emptying, postoperative pancreatic fistula, or readmissions. With POD 1 NGT removal, diet tolerance was earlier without increased NGT reinsertions. In low-risk patients, younger age, POD 1 NGT removal, and POD ≤ 3 drain removal were independent predictors of accelerated discharge. In high-risk patients, POD 1 NGT removal and POD ≤ 3 drain removal were independent predictors of accelerated discharge. CONCLUSIONS: Following iterative revisions in RSPCPs, LOS after PD decreased further without increasing readmissions, and NGTs were removed earlier without increased reinsertions. Early NGT and drain removal are modifiable practices within RSPCPs that are associated with accelerated discharge.

Robotic Completion Radical Cholecystectomy with Fluorescence Guidance.

The application of minimally invasive surgery (MIS) techniques in the treatment of hepatobiliary malignancies offers advantages of shorter length of stay, quicker functional recovery, and decreased need for postoperative opioids. However, MIS completion radical cholecystectomy for incidentally diagnosed gallbladder cancer can be challenging due to a reoperative field and lack of tactile feedback. This video demonstrates the utility of the robotic platform and highlights the ways in which it assists surgeons in overcoming these limitations. These include (1) versatile wristed instruments and excellent visualization that facilitate a thorough regional lymphadenectomy; and (2) built-in fluorescence imaging technology that can be used with intravenous indocyanine green (ICG) to confirm porta hepatis anatomy in a reoperative field. ICG pharmacokinetics enable fluorescence angiography 15-20 s after ICG injection and fluorescence cholangiography 15-20 min after ICG injection as the dye accumulates in the biliary system. Systematic and intentional application of these techniques allows for the safe performance of robotic completion radical cholecystectomy following sound oncologic principles, with excellent perioperative outcomes.

Surgical Boot Camps Increases Confidence for Residents Transitioning to Senior Responsibilities.

OBJECTIVE: Surgical boot camps enhance the confidence of medical students and surgical interns. The impact of boot camps on the confidence of post-graduate year (PGY) 2 residents is unknown. We hypothesized that a postinternship boot camp would improve the confidence of PGY-2 residents in managing their newfound responsibilities. We also hypothesized that the effect of a tailored high-impact boot camp would persist over time. DESIGN: A 2-hour boot camp at our simulation center was implemented for PGY-2 residents in 2016 and 2017. Confidence in handling boot camp scenarios was measured on a 1 to 5 Likert scale before and after the boot camp. Three-month follow-up was assessed in the 2017 cohort. PARTICIPANTS: Thirty-one PGY-2 residents (n = 16 in 2016, n = 15 in 2017) completed the boot camp. RESULTS: Residents reported increased confidence in placing central lines (p < 0.001), placing chest tubes (p = 0.01), managing emergency airways (p < 0.001), running a code (p = 0.03), and fulfilling the role of in-house senior resident (p < 0.001). Three-month follow-up in 2017 (n = 10) demonstrated no difference in confidence compared to postboot camp results. CONCLUSIONS: Boot camps can durably improve confidence in skills expected of PGY-2 residents assuming in-house senior resident responsibilities.

Intraoperative Molecular Imaging Utilizing a Folate Receptor-Targeted Near-Infrared Probe Can Identify Macroscopic Gastric Adenocarcinomas.

PURPOSE: Current methods of assessing disease burden in gastric adenocarcinoma are imperfect. Improved visualization during surgery with intraoperative molecular imaging (IMI) could improve gastric adenocarcinoma staging and guide surgical decision-making. The goal of this study was to evaluate if IMI with a folate receptor-targeted near-infrared fluorescent agent, OTL38, could identify gastric adenocarcinomas during surgery. PROCEDURES: Five patients were enrolled in an IMI clinical trial. Patients received a folate receptor-targeted near-infrared dye (OTL38) 1.5-6 h prior to surgery. During staging laparoscopy and gastric resection, IMI was utilized to identify the primary tumor and any fluorescent lymph nodes. Resected tumors were analyzed for folate receptor alpha (FRα) and CD68 expression using immunohistochemistry. Microscopic OTL38 accumulation was examined with immunofluorescence. RESULTS: Four out of five patients underwent total or subtotal gastrectomy; one had a staging laparoscopy only. All four patients who underwent gastric resection had invasive gastric adenocarcinoma; three had fluorescent tumors, mean tumor to background ratio (TBR) 4.1 ± 2.9. The one patient with a non-fluorescent tumor had a T1a tumor with two 0.4 cm tumor foci within a larger polyp. In each case with a fluorescent tumor, the fluorescence was evident from the exterior of the stomach. Two of the fluorescent tumors had modest FRα expression and no CD68 expression. One fluorescent tumor had high CD68 expression and no FRα expression. CONCLUSIONS: Intraoperative molecular imaging of gastric adenocarcinoma with OTL38 is feasible. Further studies should evaluate the clinical utility of this technique.

Comparison of a Short Versus Long Stokes Shift Near-Infrared Dye During Intraoperative Molecular Imaging.

PURPOSE: Intraoperative molecular imaging (IMI) utilizes optical dyes that accumulate within tumors to assist with detection during a cancer operation. IMI can detect disease not visualized preoperatively, as well as positive margins. However, these dyes are limited by autofluorescence, signal reflection, and photon-scatter. We hypothesize that a novel dye with a wide separation between excitation and emission spectra, SS180, would help overcome these obstacles. PROCEDURES: Two targeted molecular contrast agents, OTL38 and SS180, were selected for this study. Both dyes had the same targeting ligand to folate receptor alpha (FRα). OTL38, a well-annotated IMI agent in human trials, has a Stokes shift of 22 nm, whereas SS180, the new dye, has a Stokes shift of 129 nm. Cell lines were tested for FRα expression and incubated with dyes to demonstrate receptor-dependent binding. Cells were incubated in various concentrations of the dyes to compare dose- and time-dependent binding. Finally, cells tagged with the dyes were injected subcutaneously in a murine model to estimate tumor burden necessary to generate fluorescent signal. RESULTS: Cellular studies demonstrated that SS180 binds cells in a dose-, receptor-, and time-dependent manner and exhibits higher mean fluorescence intensities by flow cytometry when compared with OTL38 for each time point and concentration. In an in vivo flank tumor model, SS180 had a higher tumor-to-background ratio (TBR) than OTL38, though not statistically significant (p = 0.08). Ex vivo, OTL38 had a higher TBR than SS180 (p = 0.02). The subcutaneous model revealed that SS180 had a higher TBR at 5 × 106 cells than OTL38 (p = 0.05). No toxicity was observed in the animals. CONCLUSIONS: SS180 exhibits greater TBRs in vivo, but not ex vivo. These findings suggest that SS180 may have weaker fluorescence, but superior contrast. Studies in large animal models and clinical trials may better elucidate the clinical value of a long Stokes shift.

Near-infrared intraoperative imaging for minimally invasive pulmonary metastasectomy for sarcomas.

BACKGROUND: Complete pulmonary metastasectomy for sarcoma metastases provides patients an opportunity for long-term survival and possible cure. Intraoperative localization of preoperatively identified metastases and identification of occult lesions can be challenging. In this trial, we evaluated the efficacy of near-infrared (NIR) intraoperative imaging using second window indocyanine green during metastasectomy to identify known metastases and to detect occult nodules. METHODS: Thirty patients with pulmonary nodules suspicious for sarcoma metastases were enrolled in an open-label, feasibility study (NCT02280954). All patients received intravenous indocyanine green (5 mg/kg) 24 hours before metastasectomy. Patients 1 through 10 (cohort 1) underwent metastasectomy via thoracotomy to assess fluorescence patterns of nodules detected by traditional methods (preoperative imaging and intraoperative visualization/bimanual palpation). After confirming reliability within cohort 1, patients 11 through 30 (cohort 2) underwent video-assisted thoracic surgery metastasectomy with NIR imaging. RESULTS: In cohort 1, 14 out of 16 preoperatively identified pulmonary metastases (87.5%) displayed tumor fluorescence. Nonfluorescent metastases were deeper than fluorescent metastases (2.1 cm vs 1.3 cm; P = .03). Five out of 5 metastases identified during thoracotomy displayed fluorescence. NIR imaging identified 3 additional occult lesions in this cohort. In cohort 2, 33 out of 37 known pulmonary metastases (89.1%) displayed fluorescence. Nonfluorescent tumors were deeper than 2.0 cm (P = .007). NIR imaging identified 24 additional occult lesions. Of 24 occult lesions, 21 (87.5%) were confirmed metastases and the remaining 3 nodules were lymphoid aggregates. CONCLUSIONS: NIR intraoperative imaging with indocyanine green (5 mg/kg and 24 hours before surgery) localizes known sarcoma pulmonary metastases and identifies otherwise occult lesions. This approach may be a useful intraoperative adjunct to improve metastasectomy.

Intraoperative Near-infrared Imaging Can Identify Neoplasms and Aid in Real-time Margin Assessment During Pancreatic Resection.

OBJECTIVE: To determine if intraoperative near-infrared (NIR) imaging carries benefit in resection of pancreatic neoplasms. BACKGROUND: Resection of pancreatic malignancies is hindered by high rates of local and distant recurrence from positive margins and unrecognized metastases. Improved tumor visualization could improve outcomes. We hypothesized that intraoperative NIR imaging with a clinically approved optical contrast agent could serve as a useful adjunct in assessing margins and extent of disease during pancreatic resections. METHODS: Twenty patients were enrolled in an open-label clinical trial from July 2016 to May 2018. Subjects received second window indocyanine green (ICG) (2.5-5 mg/kg) 24 hours prior to pancreatic resection. NIR imaging was performed during staging laparoscopy and after pancreas mobilization in situ and following resection ex vivo. Tumor fluorescence was quantified using tumor-to-background ratio (TBR). Fluorescence at the specimen margin was compared to pathology evaluation. RESULTS: Procedures included 9 pancreaticoduodenectomies, 10 distal pancreatectomies, and 1 total pancreatectomy; 21 total specimens were obtained. Three out of 8 noninvasive tumors were fluorescent (mean TBR 2.59 ±â€Š2.57). Twelve out of 13 invasive malignancies (n = 12 pancreatic adenocarcinoma, n = 1 cholangiocarcinoma) were fluorescent (mean TBR 4.42 ±â€Š2.91). Fluorescence at the transection margin correlated with final pathologic assessment in 12 of 13 patients. Following neoadjuvant therapy, 4 of 5 tumors were fluorescent; these 4 tumors showed no treatment response on pathology assessment. One tumor had a significant treatment response and showed no fluorescence. CONCLUSIONS: Second window ICG reliably accumulates in invasive pancreatic malignancies and provides real-time feedback during pancreatectomy. NIR imaging may help to assess the response to neoadjuvant therapy.

Kinetics of indocyanine green: Optimizing tumor to normal tissue fluorescence in image-guided oral cancer surgery applications.

BACKGROUND: The study aimed to define indocyanine green (ICG) kinetics to determine the optimal ICG dose and surgical time for near-infrared fluorescence-guided oral cancer surgery. METHODS: Spectrometer and grayscale digital imaging were used to quantify the ICG kinetics in 12 patients with oral cancer. The fluorescence intensity and signal-to-background ratio (SBR) of tumor and normal tissue were tested at 1, 6, and 24 hours after ICG injection. RESULTS: The greatest contrast in the fluorescence intensity between tumor and normal tissue was observed at 6 hours (P < .01), and of three dose groups (0.5, 0.75, and 1.0 mg/kg), 0.75 mg/kg showed the highest SBR (2.06 ± 0.23) after ICG injection. CONCLUSIONS: Fluorescence quantification based on spectrometry and grayscale imaging could be effective in determining the optimal ICG dose and surgical time after ICG injection in this cohort of patients with oral cancer.

Optimization of Second Window Indocyanine Green for Intraoperative Near-Infrared Imaging of Thoracic Malignancy.

BACKGROUND: Near-infrared (NIR) imaging using the second time window of indocyanine green (ICG) allows localization of pulmonary, pleural, and mediastinal malignancies during surgery. Based on empirical evidence, we hypothesized that different histologic tumor types fluoresce optimally at different ICG doses. STUDY DESIGN: Patients with thoracic tumors biopsy-proven or suspicious for malignancy were enrolled in an NIR imaging clinical trial. Patients received a range of ICG doses 1 day before surgery: 1 mg/kg (n = 8), 2 mg/kg (n = 8), 3 mg/kg (n = 13), 4 mg/kg (n = 8), and 5 mg/kg (n = 8). Intraoperatively, NIR imaging was performed. The endpoint was to identify the highest tumor-to-background fluorescence ratio (TBR) for each tumor type at each dose. Final pathology confirmed tumor histology. RESULTS: Of 45 patients, 41 had malignancies (18 non-small cell lung cancers [NSCLC], 3 pulmonary neuroendocrine tumors, 13 thoracic metastases, 4 thymomas, 3 mesotheliomas). At doses of 4 to 5 mg/kg, the TBR from primary NSCLC vs other malignancies was no different (2.70 vs 3.21, p = 1.00). At doses of 1 to 3 mg/kg, the TBR was greater for the NSCLCs (3.19 vs 1.49, p = 0.0006). Background fluorescence from the heart or ribs was observed in 1 of 16 cases at 1 to 2 mg/kg, 5 of 13 cases at 3 mg/kg, and 14 of 16 cases at 4 to 5 mg/kg; this was a major determinant of dose optimization. CONCLUSIONS: This is the first study to demonstrate that the optimal NIR contrast agent dose varies by tumor histology. Lower dose ICG (2 to 3 mg/kg) is superior for nonprimary lung cancers, and high dose ICG (4 to 5 mg/kg) is superior for lung cancers. This will have major implications as more intraoperative imaging trials surface in other specialties, will significantly reduce costs and may facilitate wider application.

A Clinical Trial of TumorGlow to Identify Residual Disease During Pleurectomy and Decortication.

BACKGROUND: Macroscopic complete resection can improve survival in a select group of patients with malignant pleural mesothelioma. During resection, differentiating residual tumor from inflammation or scar can be challenging. This trial evaluated near-infrared (NIR) intraoperative imaging using TumorGlow (a novel NIR imaging approach utilizing high-dose indocyanine green and delayed imaging) technology to improve detection of macroscopic residual disease. METHODS: Twenty subjects were enrolled in an open-label clinical trial of NIR intraoperative imaging with TumorGlow (Indocyanine Green for Solid Tumors [NCT02280954]). Twenty-four hours before pleural biopsy or pleurectomy and decortication (P/D), patients received intravenous indocyanine green. All specimens identified during standard-of-care surgical resection and with NIR imaging underwent histopathologic profiling and correlative microscopic fluorescent tomographic evaluation. For subjects undergoing P/D (n = 13), the hemithorax was evaluated with NIR imaging during P/D to assess for residual disease. When possible, additional fluorescent lesions were resected. RESULTS: Of 203 resected specimens submitted for evaluation, indocyanine green accumulated within 113 of 113 of resected mesothelioma specimens, with a mean signal-to-background fluorescence ratio of 3.1 (SD, 2.2 to 4.8). The mean signal-to-background fluorescence ratio of benign tissues was 2.2 (SD, 1.4 to 2.4), which was significantly lower than in malignant specimens (p = 0.001). NIR imaging identified occult macroscopic residual disease in 10 of 13 subjects. A median of 5.6 resectable residual deposits per patient (range, 0 to 11 deposits per patient), with a mean size of 0.3 cm (range, 0.1 to 1.5 cm), were identified. CONCLUSIONS: TumorGlow for malignant pleural mesothelioma is safe and feasible. Excellent sensitivity allows for to reliable detection of macroscopic residual disease during cytoreductive surgical procedures.

Implications of Hospital Volume on Costs Following Esophagectomy in the United States.

BACKGROUND: With increasing focus on health care quality and cost containment, volume-based referral strategies have been proposed to improve value in high-cost procedures, such as esophagectomy. While the effect of hospital volume on outcomes has been demonstrated, our goal was to evaluate the economic consequences of volume-based referral practices for esophagectomy. METHODS: The nationwide inpatient sample (NIS) was queried for the years 2004-2013 for all patients undergoing esophagectomy. Patients were stratified by hospital volume quartile and substratified by preoperative risk and age. Clustered multivariable hierarchical logistic regression analysis was used to assess adjusted costs and mortality. RESULTS: In total, 9270 patients were clustered based on annual hospital volume quartiles of < 7, 7 to 22, 23 to 87, and > 87 esophagectomies. After stratification by patient variables, high-volume centers performed esophagectomies in high-risk patients at the same cost as low-volume centers without significant difference in resource utilization. Overall, mortality decreased across volume quartiles (lowest 8.9 versus highest 3.6%, p < 0.0001). The greatest volume-mortality differences were observed among patients aged between 70 and 80 years (lowest 12.2 versus highest 6.2%, p = 0.009). Patients with high preoperative risk also derived mortality benefits with increasing hospital volume (lowest 17.5 versus highest 11.8%, p < 0.0001). CONCLUSIONS: This study demonstrates that the mortality improvements for high-risk patients undergoing esophagectomy at high-volume centers do not come at increased costs. These results suggest that health systems should consider selectively referring high-risk patients to high-volume centers within their region.

Intraoperative fluorescence imaging in thoracic surgery.

Intraoperative fluorescence imaging (IFI) can improve real-time identification of cancer cells during an operation. Phase I clinical trials in thoracic surgery have demonstrated that IFI with second window indocyanine green (TumorGlow® ) can identify subcentimeter pulmonary nodules, anterior mediastinal masses, and mesothelioma, while the use of a folate receptor-targeted near-infrared agent, OTL38, can improve the specificity for diagnosing tumors with folate receptor expression. Here, we review the existing preclinical and clinical data on IFI in thoracic surgery.

An open label trial of folate receptor-targeted intraoperative molecular imaging to localize pulmonary squamous cell carcinomas.

BACKGROUND: Clinical applicability of folate receptor-targeted intraoperative molecular imaging (FR-IMI) has been established for surgically resectable pulmonary adenocarcinoma. A role for FR-IMI in other lung cancer histologies has not been studied. In this study, we evaluate feasibility of FR-IMI in patients undergoing pulmonary resection for squamous cell carcinomas (SCCs). METHODS: In a human clinical trial (NCT02602119), twelve subjects with pulmonary SCCs underwent FR-IMI with a near-infrared contrast agent that targets the folate receptor-α (FRα), OTL38. Near-infrared signal from tumors and benign lung was quantified to calculate tumor-to-background ratios (TBR). Folate receptor-alpha expression was characterized, and histopathologic correlative analyses were performed to evaluate patterns of OTL38 accumulation. An exploratory analysis was performed to determine patient and histopathologic variables that predict tumor fluorescence. RESULTS: 9 of 13 SCCs (in 9 of 12 of subjects) displayed intraoperative fluorescence upon NIR evaluation (median TBR, 3.9). OTL38 accumulated within SCCs in a FRα-dependent manner. FR-IMI was reliable in localizing nodules as small as 1.1 cm, and prevented conversion to thoracotomy for nodule localization in three subjects. Upon evaluation of patient and histopathologic variables, in situ fluorescence was associated with distance from the pleural surface, and was independent of alternative variables including tumor size and metabolic activity. CONCLUSIONS: This work demonstrates that FR-IMI is potentially feasible in 70% of SCC patients, and that molecular imaging can improve localization during minimally invasive pulmonary resection. These findings complement previous data demonstrating that ∼98% of pulmonary adenocarcinomas are localized during FR-IMI and suggest broad applicability for NSCLC patients undergoing resection.

Near-infrared intraoperative imaging during resection of an anterior mediastinal soft tissue sarcoma.

Sarcomas are rare malignancies that are generally treated with multimodal therapy protocols incorporating complete local resection, chemotherapy and radiation. Unfortunately, even with this aggressive approach, local recurrences are common. Near-infrared intraoperative imaging is a novel technology that provides real-time visual feedback that can improve identification of disease during resection. The presented study describes utilization of a near-infrared agent (indocyanine green) during resection of an anterior mediastinal sarcoma. Real-time fluorescent feedback provided visual information that helped the surgeon during tumor localization, margin assessment and dissection from mediastinal structures. This rapidly evolving technology may prove useful in patients with primary sarcomas arising from other locations or with other mediastinal neoplasms.

A Phase I Clinical Trial of Targeted Intraoperative Molecular Imaging for Pulmonary Adenocarcinomas.

BACKGROUND: Intraoperative identification of pulmonary nodules, particularly small lesions, can be challenging. We hypothesize that folate receptor-targeted intraoperative molecular imagining can be safe and improve localization of pulmonary nodules during resection. METHODS: Twenty subjects with biopsy-proven pulmonary adenocarcinomas were enrolled in a phase I clinical trial to test the safety and feasibility of OTL38, a novel folate receptor-α (FRα) targeted optical contrast agent. During resection, tumors were imaged in situ and ex vivo and fluorescence was quantified. Resected specimens were analyzed to confirm diagnosis, and immunohistochemistry was utilized to quantify FRα expression. A multivariate analysis using clinical and tumor data was performed to determine variables impacting tumor fluorescence. RESULTS: Of the 20 subjects, three grade I adverse events were observed: all transient nausea/abdominal pain. All symptoms resolved after completing the infusion. Sixteen of 20 subjects (80%) had tumors with in situ fluorescence with a mean tumor-to-background fluorescence level of 2.9 (interquartile range, 2.1 to 4.2). The remaining 4 subjects' tumors fluoresced ex vivo. In situ fluorescence was dependent on depth from the pleural surface. Four subcentimeter nodules not identified on preoperative imaging were detected with intraoperative imaging. CONCLUSIONS: This phase I trial provides preliminary evidence suggesting that folate receptor-targeted molecular imaging with OTL38 is safe, with tolerable grade I toxicity. These data also suggest that OTL38 accumulates in known lung cancers and may improve identification of synchronous malignancies. Our group is initiating a five-center, phase II study to better understand the clinical implications of intraoperative molecular imaging using OTL38.

Surgical Management of Early-Stage Esophageal Adenocarcinoma Based on Lymph Node Metastasis Risk.

BACKGROUND: In early-stage esophageal adenocarcinoma (EAC), esophagectomy improves staging but also increases mortality compared with endoscopic resection. Our objective was to quantify esophagectomy mortality and lymph node metastasis (LNM) risk in early-stage EAC to improve surgical treatment allocation. METHODS: We identified National Cancer Database (2004-2014) patients with nonmetastatic, Tis, T1a, or T1b EAC who had primary surgical resection and microscopic examination of at least 15 lymph nodes. Univariate and multivariable logistic regression identified predictors of LNM. Cox regression identified predictors of death. The Kaplan-Meier method predicted overall survival (OS). RESULTS: In 782 patients, LNM rates were: all patients 13.8%, Tis 0%, T1a 3.6%, T1b 23.4%. Independent predictors of LNM were submucosal invasion, lymphovascular invasion (LVI), decreasing differentiation, and tumor size ≥ 2 cm (P < 0.05). For T1a tumors with poor differentiation or size ≥ 2 cm, LNM rates were 10.2 and 6.7%, respectively; 90-day mortality was 3.1%. The LNM rate in well differentiated T1b tumors < 2 cm was 4.2%; 90-day mortality was 6.0%. Estimated 5-year OS was 80.2% versus 64.4% (T1a vs. T1b). LNM increased risk of death for T1a (hazard ratio [HR] 8.52, 95% confidence interval [CI] 3.13-23.22, P < 0.001) and T1b tumors (HR 2.52, 95% CI 1.59-4.00, P < 0.001). CONCLUSIONS: In T1a EAC with poor differentiation or size ≥ 2 cm, esophagectomy should be considered, whereas in T1b EAC with low-risk features (well-differentiated T1b EAC < 2 cm without LVI), endoscopic resection may be sufficient. Treatment guidelines for early-stage EAC should include all high-risk tumor features for LNM and stage-specific esophagectomy mortality.