Search

Implant-based breast reconstruction infections often require implant explantation. Whereas some plastic surgeons pursue autologous reconstruction following the first implant-based breast reconstruction failure caused by infection, others argue that a second attempt is acceptable.The authors conducted a retrospective study of patients who underwent a second reconstruction attempt with implant-based or free flap breast reconstruction following explantation because of infection between 2006 and 2019. Surgical and patient-reported outcomes were compared between the two groups.A total of 6093 implant-based breast reconstructions were performed during the study period, of which 130 breasts met our inclusion criteria [implant-based, n = 86 (66 percent); free flap, n = 44 (34 percent)]. No significant differences in rates of overall (25 percent versus 36 percent; p = 0.2) or major (20 percent versus 21 percent; p = 0.95) complications were identified between the free flap and implant-based cohorts, respectively. Implant-based breast reconstruction patients were more likely to experience a second infection (27 percent versus 2 percent; p = 0.0007) and reconstruction failure (21 percent versus 5 percent; p = 0.019). Among irradiated patients, reconstruction failure was reported in 44 percent of the implant-based and 7 percent of the free flap cohorts (p = 0.02). Free flap patients reported significantly higher scores for Satisfaction with Breasts (73.7 ± 20.1 versus 48.5 ± 27.9; p = 0.0046).Following implant-based breast reconstruction explantation because of infection, implant-based and free flap breast reconstruction had similar rates of overall and major complications; however, implant-based breast reconstruction had considerably higher rates of infection and reconstructive failures and lower patient-reported scores for Satisfaction with Breasts. Given the high rates of implant-based breast reconstruction failure in patients with prior radiotherapy and infection-based failure, plastic surgeons should strongly consider autologous reconstruction in this patient population.Therapeutic, III.

Infection is a dreaded complication of implant-based breast reconstruction. There is a paucity of literature on the outcomes of a secondary reconstruction after infected implant-based breast reconstruction explantation.The authors conducted a retrospective study of patients who underwent a second implant-based breast reconstruction following a failed infected device between January of 2006 and December of 2019. Surgical and patient-reported outcomes (BREAST-Q) were collected and analyzed.A total of 6093 implant-based breast reconstructions were performed during the study period, 298 (5 percent) of which involved device removal because of infection. Eighty-three patients ultimately received 92 second-attempt breast implants. Thirty-six percent of cases developed at least one postoperative complication, with infection [23 breasts (25 percent)] being the most common. Compared with first-attempt implant-based breast reconstruction, we found significantly higher infection rates among second-attempt cases (9 percent and 21 percent, respectively; p = 0.0008). Patient-reported satisfaction with the breast and sexual well-being were lower after second-attempt than after first-attempt implant-based breast reconstruction ( p = 0.018 and p = 0.002, respectively) reported in the literature. Mean follow-up was 41 ± 35 months. If we exclude patients with prior radiation therapy and those who received device exchange, the success rate is 88 percent.It is reasonable to offer women second-attempt implant-based breast reconstruction after explantation because of infection. However, this patient population has a higher infection and explantation rate and lower patient-reported satisfaction than patients undergoing first-attempt implant-based breast reconstruction. Because of these increased surgical risks and elevated complication rates, patients must be given reasonable expectations during preoperative discussions and when providing informed consent for second-attempt implant-based breast reconstruction.Risk, III.

Background Robotically assisted latissimus dorsi harvest permits harvest of the latissimus dorsi muscle without a back incision, as compared to the traditional open technique. The authors hypothesized that robotic harvest has lower donor-site complication rates, decreased opioid requirements, and a shorter length of stay than the traditional open technique. Methods A retrospective review was performed of all consecutive pedicled latissimus dorsi flaps for breast reconstruction between 2011 and 2015. All procedures were conducted by two surgeons who performed both robotic and open cases. Results Fifty-two patients were identified; 25 underwent robotically assisted latissimus dorsi harvest and 27 underwent the open technique. Demographic data between the two groups were similar. Median length of stay for robotic harvest was shorter than that for the traditional technique (2 days versus 3 days; p = 0.031). Postoperative morphine requirement was less in the robotic compared to the traditional technique, but the difference was not significant (158 mg versus 184 mg; p = 0.826). Seroma rate was higher in the robotic group (16 percent versus 0 percent; p = 0.034). The mean duration of surgery was longer in the robotic cohort (388 minutes versus 311 minutes; p = 0.002). Conclusions This study demonstrates robotically assisted latissimus dorsi harvest as an effective alternative to the traditional open technique in select patients. Advantages of robotic harvest include no back scar, a shorter length of stay, and lower opioid requirements, although the difference was not significant; disadvantages include longer operative time and a higher seroma rate. Clinical question/level of evidence Therapeutic, III.

Background Type 1 internal hemipelvectomies and total sacrectomies cause significant biomechanical instability, demanding pelvic ring reconstruction for ambulation and torso support. Previously described methods include autografts, allografts, and implants, commonly with poor long-term outcomes. The authors hypothesized that the free fibula flap for spinopelvic reconstruction is safe and effective, and associated with a high bony union rate and superior functional outcomes. Methods The authors performed a retrospective review of all patients who underwent free fibula flap surgery after internal hemipelvectomy or total sacrectomy at M. D. Anderson Cancer Center from 2003 to 2018. The primary outcome was radiographic evidence of bony union. Secondary outcomes included surgical-site occurrence and lower extremity function. Univariate and multivariate logistic regression analyses were performed. Results Forty-seven patients were included (internal hemipelvectomy, n = 38; total sacrectomy, n = 9). The mean follow-up was 3.3 years and the most common abnormality was chondrosarcoma (30.4 percent). The nonunion rate was 9.7 percent and the surgical-site occurrence rate was 34 percent; there were no flap losses. Greater age was significantly associated with nonunion (OR, 1.1; 95 percent CI, 1 to 1.2; p = 0.003), whereas obesity was the only independent predictor of surgical-site occurrence (OR, 9.2; 95 percent CI, 1.2 to 71.3; p = 0.03). Functional metrics approached those of adult norms by 3 years postoperatively. Compared to internal hemipelvectomy patients, patients undergoing total sacrectomy had more comorbidities, a higher complication rate, and a worse functional outcome. Conclusion The free fibula flap for spinopelvic reconstruction is safe and effective, and is associated with a high bony union rate and superior functional outcomes. Clinical question/level of evidence Risk, III.

Reducing complications while controlling costs is a central tenet of value-based health care. Bilateral microvascular breast reconstruction is a long operation with a relatively high complication rate. Using a two-surgeon team has been shown to improve safety in bilateral microvascular breast reconstruction; however, its impact on cost and efficiency has not been robustly studied. The authors hypothesized that a cosurgeon for bilateral microvascular breast reconstruction is safe, effective, and associated with reduced costs.The authors retrospectively reviewed all patients who underwent bilateral microvascular breast reconstruction with either a single surgeon or surgeon/cosurgeon team over an 18-month period. Charges were converted to costs using the authors' institutional cost-to-charge ratio. Surgeon opportunity costs were estimated using time-driven activity-based costing. Propensity scoring controlled for baseline characteristics between the two groups. A locally weighted logistic regression model analyzed the cosurgeon's impact on outcomes and costs.The authors included 150 bilateral microvascular breast reconstructions (60 single-surgeon and 90 surgeon/cosurgeon reconstructions) with a median follow-up of 15 months. After matching, the presence of a cosurgeon was associated with a significantly reduced mean operative duration (change in operative duration, -107 minutes; p0.001) and cost (change in total cost, -$1101.50; p0.001), which was even more pronounced when surgeon/cosurgeon teams worked together frequently (change in operative duration, -132 minutes; change in total cost, -$1389; p = 0.007). The weighted logistic regression models identified that a cosurgeon was protective against breast-site complications and trended toward reduced overall and major complication rates.The practice of using a of cosurgeon appears to be associated with reduced costs and improved outcomes, thereby potentially adding value to bilateral microvascular breast reconstruction.Therapeutic, III.

Background A 56-year-old woman presented with an extensive sarcoma requiring nearly total back resection. She had limited donor sites for reconstruction because of a previous laparotomy, but presented with a significantly larger, identical twin. Cancer has traditionally been considered a contraindication for vascularized composite allotransplantation; however, immunosuppression is potentially avoidable between monozygotic twins. Methods A preoperative genetic workup revealed 10/10 human leukocyte antigen homozygosity. Despite substantial phenotypic divergence in size and facial features, the sisters were genotypically identical. A two-stage, double deep inferior epigastric perforator transplant was planned for delayed reconstruction. At the first stage following the resection, an arteriovenous loop was performed to provide recipient vasculature to the back. At a second stage, the transplantation was performed. In addition, bilateral lumbar artery perforator flaps were created to reduce the length of the defect. Intraoperative steroid bolus and a short taper alone were used for immunosuppression. Results The resection resulted in a 22 × 29-cm specimen down to the spine. After a 4-day interval for permanent pathologic evaluation, the transplant was successfully transferred between twins. Two arteries and six veins were anastomosed to establish perfusion. Postoperatively, there have been no episodes of rejection or flap compromise at last follow-up (>36 months). Conclusions This case represents one of the few vascularized composite allotransplantations between monozygotic twins, and the only reported successful vascularized composite allotransplantation for a recurrent cancer diagnosis. Oncologic safety depended on 100 percent histocompatibility to avoid immunosuppression. Limited patient donor sites precluded total autologous coverage, and a substantial size discrepancy between the twins favored a transplant.

Background Advancements in three-dimensional (3D) printing have enabled production of patient-specific guides to aid perforator mapping and pedicle dissection during abdominal flap harvest. We present our early experience using this tool to navigate deep inferior epigastric artery (DIEA) topography and evaluate its impact on operative efficiency and clinical outcomes. Patients and Methods Between January 2013 and December 2018, a total of 50 women underwent computed tomographic angiography (CTA)-guided perforator mapping prior to abdominal flap breast reconstruction, with (n = 9) and without (n = 41) 3D-printed vascular modeling (3DVM). Models were assessed for their accuracy in identifying perforator location and source-vessel anatomy, as determined by operative findings from 18 hemi-abdomens. The margin of error (MOE) for perforator localization using 3DVM was calculated and compared with CTA-derived measurements for the same patients. Flap harvest times, outcomes, and complications for patients who were preoperatively mapped using 3DVM versus CTA alone were analyzed. Results Overall, complete concordance was observed between 3DVM and operative findings with regards to perforator number, source-vessel origin, and DIEA branching pattern. By comparison, CTA interpretation of these parameters inaccurately identified branching pattern and perforator source-vessel origin in 28 and 33% of hemi-abdomens, respectively (p = 0.045 and p = 0.02). Compared with operative measurements, the average MOE for perforator localization using 3DVM was significantly lower than that obtained from CTA alone (0.81 vs. 8.71 mm, p Conclusion The results of this study support the accuracy of 3DVM for identifying DIEA topography and perforator location. Application of this technology may translate to enhanced operative efficiency and fewer perfusion-related complications for patients undergoing abdominal free flap breast reconstruction.

BACKGROUND Although the profunda artery perforator flap has gained popularity in breast reconstruction, it has not been well described for reconstruction of head and neck defects. The authors report their experience with free profunda artery perforator flaps in postoncologic head and neck reconstruction. METHODS A retrospective review of all free profunda artery perforator flaps used for head and neck reconstruction from 2016 to 2019 was performed. RESULTS Overall, 61 profunda artery perforator flap reconstructions were performed: 45 single independent flaps, 12 in conjunction with a second free flap, and four in combination with two other free flaps. The profunda artery perforator flaps were most commonly used for reconstruction of the tongue (n = 19), cheek (n = 11), parotid (n = 10), and maxilla (n = 6). The profunda artery perforator flaps averaged 7.1 × 12.1 × 1.9 cm, with a mean pedicle length of 11.5 cm. The A, B, and C perforators were located at mean distances of 7.4 cm (range, 4 to 11.5 cm), 11.7 cm (range, 8 to 18 cm), and 16.1 cm (range, 14 to 20.5 cm) from the pubic tubercle along the axis of the adductor longus muscle and 7.9 cm (range, 7 to 11cm), 7.6 cm (range, 7 to 15.5 cm), and 7.2 cm (range, 6 to 16 cm) posterior and perpendicular to the axis. There were three partial flap losses. Eight patients (13 percent) had recipient-site complications necessitating operative intervention: four for vascular compromise of the profunda artery perforator flap, two for hematoma evacuation, and two for infection. Donor-site complications were noted in seven patients (11 percent), two of whom required operative intervention. CONCLUSIONS The profunda artery perforator flap is a versatile and reliable flap with consistent anatomy and a low complication rate. The profunda artery perforator flap seems to be a reasonable alternative for reconstruction of head and neck defects. CLINICAL QUESTION/LEVEL OF EVIDENCE Therapeutic, IV.

Background Discovering alternatives to workhorse flaps that have more consistent anatomy and lower donor-site morbidity has become a focus of reconstructive surgery research. This study provides a simplified approach to profunda artery perforator flap design and harvest based on reliable anatomical landmarks. Methods A retrospective review was conducted of 70 patients who underwent 83 profunda artery perforator flap reconstructions for postoncologic defects from 2016 to 2018. The authors recorded and analyzed the profunda artery perforator flap sizes and clinical applications, the numbers and locations of the perforators, and the patient outcomes. Results Most of the profunda artery perforator flaps were for head and neck [46 patients (65.7 percent)] and breast [21 patients (30 percent)] reconstructions. Flaps were most commonly based on perforator A (33.7 percent) and perforator B (33.7 percent), followed by perforators B and C combined (18.1 percent). Perforators were located a mean of 7.5 cm (perforator A), 12.7 cm (B), and 17.6 cm (C) distal to the pubic tubercle parallel to the axis between the pubic tubercle and the medial femoral condyle and 7.9 cm (A), 7.3 cm (B), and 6.1 cm (C) posterior from the axis. There was no flap loss. One patient underwent successful salvage surgery after arterial flap thrombosis. Eight patients (9.6 percent) developed superficial wound dehiscence that was managed conservatively. Conclusions Perforator mapping demonstrated consistent anatomical locations of sizeable profunda artery perforators in the inner thigh. Along with its consistent and robust vascular anatomy and minimal donor-site morbidity, the profunda artery perforator flap's volume and pliability make it a reliable option for soft-tissue reconstruction. Clinical question/level of evidence Therapeutic, IV.

Contralateral prophylactic mastectomy (CPM) is increasingly performed in average-risk patients despite the lack of survival benefit. In an era of heightened awareness of healthcare costs, we sought to determine the impact of CPM on financial toxicity in breast cancer. A single-institution propensity-matched analysis of female patients who underwent unilateral mastectomy (UM) with or without CPM for breast cancer over an 18-month period. Patients with a history of genetic predisposition or bilateral cancer were excluded. The validated Comprehensive Score for financial Toxicity (COST) evaluated financial toxicity among participants. Multivariable regression analysis evaluated the relationship between CPM and financial toxicity. Relevant domains of the Breast Q and SF12 instruments were examined as secondary outcomes. Sensitivity analysis was performed using propensity-weighting to examine robustness of results and increase our sample size. Overall, 104 patients were identified, equally distributed across UM and CPM. CPM was not associated with financial toxicity, as evidenced by comparable COST scores (adjusted difference, 1.53 [− 3.24 to 6.29]). Minor complications were significantly lower in UM patients (UM, 8%; CPM, 31%). CPM was associated with significantly higher Breast Q psychosocial well-being score (adjusted difference, 10.58 [1.34 to 19.83]). BREAST Q surgeon satisfaction, SF12 mental and physical component scores were comparable. Similar results were noted on sensitivity analysis involving 194 patients. Choice for CPM was associated with higher minor complications, but led to improved psychosocial well-being without a higher degree of patient-reported financial toxicity. Prospective studies are needed to discern the influence of CPM on the incidence and trajectory of financial toxicity.

Although diabetes mellitus (DM) is a known risk factor for surgical complications in general, there is little published evidence to establish such an effect among patients undergoing breast reconstruction (BR). The purpose of this study was to assess the impact of DM on complications in patients undergoing postmastectomy BR.Patients who underwent BR at our institution from November 2002 to November 2012 were identified. Clinical and demographic data of patients with type 1 or type 2 DM were reviewed. Complications occurring within 60 days of surgery were compared between diabetic and nondiabetic patients for both autologous and nonautologous reconstruction types.A total of 1371 BR were performed in 1035 patients. There were 877 (64.0%) autologous reconstructions and 494 (36.0%) implant-based reconstructions. Patients with DM (n = 64) had significantly higher preoperative blood glucose levels (137.5 vs 100.1, P0.05). Postoperatively, DM patients reconstructed with implants had a significantly higher incidence of delayed wound healing (22.2% vs 9.7%; P = 0.04). This was not observed in patients with DM reconstructed with autologous tissue (7.4% vs 6.6%; P = 0.70). Diabetic patients had a significantly higher incidence of hypertension and were older than nondiabetic patients. To control for these variables and other potential confounders, multiple logistic regression analysis was performed. Again, diabetic patients had a significantly higher incidence of delayed wound healing following implant-based reconstruction (odds ratio, 2.52, 95% confidence interval = 1.2-6.2) but not autologous reconstruction (odds ratio, 0.97; 95% confidence interval = 0.2-4.6).Diabetes heightens the risk of wound healing complications among patients undergoing implant-based reconstruction.

Postoperative hematoma and venous congestion after free tissue transfer may occur independently or concurrently. We aimed to explore the association between these two events.All free flap reconstructions for head and neck (HN) and breast from a single institution between 2004 and 2014 were retrospectively reviewed for reoperation for venous congestion and/or hematoma.There were 2985 free flap cases for HN reconstruction and 2345 cases for breast reconstruction. In HN, 100 patients developed a hematoma (3.4%) and 84 patients developed venous congestion (2.8%). The prevalence of hematoma was 17.8% and 2.9% in the presence and absence of congestion, respectively (p0.001). Among the 15 patients who had both hematoma and venous congestion were separate events that occurred from 1 to 9 days apart in 8 patients. Hematoma caused the compression of the pedicle vein in 4 patients, while venous congestion possibly caused hematoma in 3 patients. In breast, 56 patients developed a hematoma (2.4%) and 64 patients developed venous congestion (2.7%). The prevalence of hematoma was 12.5% and 2.1% in the presence and absence of congestion, respectively (p0.001). In the 8 patients who developed both, hematoma and congestion were separate events in 4 patients. Venous congestion caused hematoma in 3 patients, and hematoma caused venous congestion in 1 patient.Although postoperative hematoma and venous congestion often present concurrently, most events are not causally associated. When related, however, venous congestion leading to hematoma is more common in breast reconstruction, while hematoma preceding venous congestion is more common in HN reconstruction.

A recent survey of plastic surgeons showed that the majority prescribed prophylactic antibiotics after hospital discharge for breast reconstruction. There is no clinical evidence that this practice reduces surgical site infection (SSI) after immediate tissue expander breast reconstruction. Furthermore, multiple studies have suggested that current antibiotic choices may not be appropriately covering the causative organisms of SSI.An institutional breast reconstruction database from January 2005 to December 2011 was queried to identify patients undergoing immediate tissue expander reconstruction of the breast. The bacteriology of the infection, prophylactic and empiric antibiotic use, and antibiotic sensitivities were analyzed.In 557 cases of immediate tissue expander breast reconstruction performed in 378 patients, SSIs were diagnosed in 50 (9.0%) cases. Two hundred patients were given oral antibiotics at discharge; 178 did not receive antibiotics. Surgical site infection developed in 12.0% of patients given oral antibiotics and in 13.5% of those not receiving antibiotics (P = 0.67). Wound culture data were obtained in 34 SSIs. Twenty-nine had positive cultures. The most common offending organisms were methicillin-sensitive (11) and methicillin-resistant (6) Staphylococcus aureus. Despite increased use of postoperative prophylaxis over the years, SSI incidence remained unchanged. However, trends toward increased resistance of SSI organisms to the preoperative and postoperative prophylaxis agents were observed. When first-generation cephalosporins were used as prophylaxis, SSI organisms showed resistance rates of 20.5% (preoperative cefazolin) and 54.5% (postoperative cephalexin).Administration of extended prophylactic antibiotics does not reduce overall risk of SSI after expander-based breast reconstruction but may influence antibiotic resistance patterns when infections occur. The organisms most commonly responsible for SSI are often resistant to cefazolin.

Prior breast augmentation in patients desiring post-mastectomy reconstruction provides a unique opportunity for capsular preservation and immediate, single-step implant reconstruction. We report a case series of a single-surgeon experience with immediate implant reconstruction after skin-sparing mastectomy in patients with prior subpectoral augmentation. Final implant volumes, complications, and outcomes were examined. Twenty patients (15 bilateral, total 35 breasts) were included. Eighteen (90%) patients were treated for cancer. Mean augmentation-to-reconstruction interval was 9 years (range, 3-19 years). Mean patient age was 45.1 years (range, 37-64 years). Eight patients (40%) received postoperative chemotherapy and two (10%) radiation. Mean mastectomy weight was 321 g. Mean weight of the implants removed was 346 g. Mean volume of new implants was 487 mL. All patients underwent capsulotomy (100% superior, 85% medial, 30% inferior, 5% lateral). Mean operative time was less than 1 hour for bilateral reconstruction. With average follow-up of 25.6 months, 2 patients were re-operated on for asymmetry (implant malposition, synmastia). Thirty-day complications included 1 implant loss due to infection, 1 drain placement with implant salvage, 1 hematoma requiring evacuation, and 1 cellulitis treated with antibiotics. There were no late complications and no capsular contractures. None have required further oncologic surgery. No cancer recurrences have been detected. In patients who desire prosthetic reconstruction similar to their original submuscular augmentation, capsule preservation and implant replacement with a larger prosthetic inserted within the old capsule is safe, fast, and aesthetically pleasing without compromising oncologic principles.

As compared with open distal pancreatectomy, laparoscopic distal pancreatectomy (LDP) is associated with lower morbidity and shorter hospital stays. Existing reports do not elucidate trends in patient selection, technique, and outcomes over time. We aimed to determine outcomes after LDP at a specialized center, analyze trends of patient selection and operative technique, and validate a complication risk score (CRS).Patients undergoing LDP between January 2000 and January 2011 were identified and divided into 2 equal groups to represent our early and recent experiences. Demographics, tumor characteristics, operative technique, and perioperative outcomes were examined and compared between groups. A CRS was calculated for the entire cohort and examined against observed outcomes.A total of 132 LDPs were attempted, of which 8 (6.1%) were converted to open procedures. Thirty-day overall and major complication rates were 43.2% and 12.9%, respectively, with mortality1%. Pancreatic fistulas occurred in 28 (21%) patients, of which 14 (11%) were clinically significant. Recent LDPs (n = 66) included patients with increasingly severe comorbidities (Charlson scores2, 40.9% vs 16.7%, p = 0.003), more proximal tumors (74.2% vs 26.2%, p0.001), more extended resections (10.6 vs 8.3 cm, p0.001), shorter operative times (141 vs 172 minutes, p = 0.007), and less frequent use of a hand port (25.8% vs 66.6%, p0.001). No significant differences were found in perioperative outcomes between the groups. As compared with the hand access technique, the total laparoscopic approach was associated with shorter hospital stays (5.3 vs 6.8 days, p = 0.032). Increasing CRS was associated with longer operative time, significant fistulas, wound infections, blood transfusions, major complications, ICU readmissions, and rehospitalizations.This large, single-institution series demonstrates that despite a shift in patient selection to sicker patients with more proximal tumors, similar perioperative outcomes can be achieved with laparoscopic distal pancreatectomy. The CRS appears to be a reliable preoperative assessment tool for assessing other adverse perioperative outcomes in addition to predicting overall complications and fistulas as originally published.

Background The head and neck have a rich lymphatic drainage and complex anatomy, which complicate sentinel lymph node (SLN) biopsy for melanoma. The incidence of regional recurrence after a negative SLN biopsy has been shown to be higher than that at other sites. Compounding factors in this scenario were analyzed to determine their impact on both SLN status and survival. Methods A retrospective review of a prospective database of 315 patients who underwent SLN biopsy for head and neck melanoma from 1994 to 2009 was performed. A false-negative SLN biopsy was defined as a regional recurrence in a previously mapped nodal basin. Results In all, 267 patients (84.8%) were SLN negative (SLN−) and 48 patients (15.2%) were SLN positive (SLN+). The false-negative SLN biopsy occurred in 17 patients (6.4%). The mean follow-up was 37.6 months (3–152 months). The false-negative SLN and SLN+ patients were similar with respect to patient age and sex and primary melanoma tumor thickness and ulceration. The site of the primary melanoma and the lymphatic drainage patterns did not influence the false-negative biopsy rate. The mean survival was SLN− 119.3 months (95% confidence interval [CI], 106.1–132.5, p < .001), SLN+ 73.4 months (95% CI, 52.3–94.4), and the false-negative SLN 70.7 months (95% CI, 54.2–87.1). Conclusions The risk of a false-negative SLN biopsy in head and neck melanoma is independent of primary site and lymphatic drainage pattern. Patients with head and neck melanoma who have a regional recurrence after a negative SLN biopsy do not have a worse survival than that of patients who are initially SLN positive. © 2012 Wiley Periodicals, Inc. Head Neck, 2013

hpb_432 228..235 Background: The effect of diabetes on survival after resection pancreatic ductal carcinoma (PDAC) is unclear. The present study was undertaken to determine whether pre-operative diabetes has any predictive value for survival. Methods: A retrospective review from seven centres was performed. Metabolic factors, tumour char- acteristics and outcomes of patients undergoing resection for PDAC were collected. Univariate and multivariable analyses were performed to determine factors associated with disease-free (DFS) and overall survival (OS). Results: Of the 509 patients in the present study, 31.2% had diabetes. Scoring systems were devised to predict OS and DFS based on a training set (n = 245) and were subsequently tested on an independent set (n = 264). Pre-operative diabetes (P 2c m (P = 0.001), metastatic nodal ratio >0.1 (P < 0.001) and R1 margin (P < 0.001) all correlated with DFS and OS on univariate analysis. Scoring systems were devised based on multivariable analysis of the above factors. Diabetes and the metastatic nodal ratio were the most important factors in each system, earning two points for OS and four points for DFS. These scoring systems significantly correlated with both DFS (P < 0.001) and OS (P < 0.001). Conclusion: Pre-operative diabetes status provides useful information that can help to stratify patients in terms of predicted post-operative OS and DFS.

Background Use of prosthetic grafts for reconstruction after portal vein (PV) resection during pancreaticoduodenectomy is controversial. We examined outcomes in patients who underwent vein reconstruction using polytetrafluoroethylene (PTFE). Study Design Review of prospectively maintained databases at 3 centers identified all patients who underwent pancreaticoduodenectomy (PD) with vein resection and reconstruction using PTFE grafts between 1994 and 2009. Patient, operative, and outcomes variables were studied. Graft patency and survival were assessed using the Kaplan-Meier technique. Results Thirty-three patients underwent segmental vein resection with interposition PTFE graft reconstruction. Median age was 67 years; median Eastern Cooperative Oncology Group score was 1. Most operations were performed for pancreatic adenocarcinoma (n = 28, 85%); 96% were T3 lesions or greater. Standard PD was performed in 12 (36%) patients, pylorus-preservation in 17 (52%), and total pancreatectomy in 4 (12%). Combined resection of portal and superior mesenteric veins (SMV) was required in 49%, with resection isolated to PV in 12% and SMV in 39%. Splenic vein ligation was necessary in 30%. Median graft diameter was 12 mm (range 8 to 20 mm), with the majority being ring-enforced (73%). Median operative and vascular clamp times were 463 and 41 minutes, respectively, with median blood loss of 1,500 mL. The negative margin rate was 64%. Overall morbidity rate was 46%, and 30-day mortality was 6%. No patients developed irreversible hepatic necrosis or graft infection. Pancreatic fistulas occurred in 3 (9.1%). With mean follow-up of 14 months, overall graft patency was 76%. Estimated median duration of graft patency was 21 months. Median survival was 12 months for pancreatic adenocarcinoma. Conclusions With careful patient selection, PTFE graft reconstruction of resected PV/SMV during pancreaticoduodenectomy is possible with minimal risk of hepatic necrosis or graft infection. Comparison studies to primary anastomosis and autologous vein reconstruction are necessary.

BACKGROUND: Management of neuroendocrine tumor liver metastasis (NELM) remains controversial, with some advocating an aggressive surgical approach while others have adopted a more conservative strategy. We sought to define the efficacy of the surgical management of NELM in a large multicenter international cohort of patients. METHODS: We identified 339 patients who underwent surgical management for NELM from 1985 to 2009 from an international database of eight major hepatobiliary centers. Relevant clinicopathologic data were assessed using Kaplan-Meier and Cox regression models. RESULTS: Most patients had a pancreatic (40%) or small bowel (25%) neuroendocrine tumor (NET) primary. The majority of patients (60%) had bilateral liver disease. At surgery, 78% of patients underwent hepatic resection, 3% ablation alone, and 19% resection + ablation. Major hepatectomy was performed in 45% of patients, and 14% underwent a second liver operation. Carcinoid was the most common NET histological subtype (53%). Median survival was 125 months, with overall 5- and 10-year survival of 74%, and 51%, respectively. Disease recurred in 94% of patients at 5 years. Patients with hormonally functional NET who had R0/R1 resection benefited the most from surgery (P = 0.01). On multivariate analyses, synchronous disease [hazard ratio (HR) = 1.9], nonfunctional NET hormonal status (HR = 2.0), and extrahepatic disease (HR = 3.0) remained predictive of worse survival (all P < 0.05). CONCLUSIONS: Liver-directed surgery for NELM is associated with prolonged survival; however, the majority of patients will develop recurrent disease. Patients with hormonally functional hepatic metastasis without prior extrahepatic or synchronous disease derive the greatest survival benefit from surgical management.

Background As compared with open distal pancreatectomy (ODP), laparoscopic distal pancreatectomy (LDP) affords improved perioperative outcomes. The role of LDP for patients with pancreatic ductal adenocarcinoma (PDAC) is not defined. Study Design Records from patients undergoing distal pancreatectomy (DP) for PDAC from 2000 to 2008 from 9 academic medical centers were reviewed. Short-term (node harvest and margin status) and long-term (survival) cancer outcomes were assessed. A 3:1 matched analysis was performed for ODP and LDP cases using age, American Society of Anesthesiologists (ASA) class, and tumor size. Results There were 212 patients who underwent DP for PDAC; 23 (11%) of these were approached laparoscopically. For all 212 patients, 56 (26%) had positive margins. The mean number of nodes (± SD) examined was 12.6 ±8.4 and 114 patients (54%) had at least 1 positive node. Median overall survival was 16 months. In the matched analysis there were no significant differences in positive margin rates, number of nodes examined, number of patients with at least 1 positive node, or overall survival. Logistic regression for all 212 patients demonstrated that advanced age, larger tumors, positive margins, and node positive disease were independently associated with worse survival; however, method of resection (ODP vs. LDP) was not. Hospital stay was 2 days shorter in the matched comparison, which approached significance (LDP, 7.4 days vs. ODP, 9.4 days, p=0.06). Conclusions LDP provides similar short- and long-term oncologic outcomes as compared with OD, with potentially shorter hospital stay. These results suggest that LDP is an acceptable approach for resection of PDAC of the left pancreas in selected patients.

To assess clinicopathologic features and postresection survival of diabetes mellitus (DM)-associated pancreatic ductal adenocarcinoma (PDAC).Records of resected PDAC patients from 2000 to 2007 were reviewed. DM was classified as new-onset (24 months before PDAC) or longstanding (or =24 months). Clinicopathologic features were compared by univariate and multivariate analyses. Survival was assessed by Kaplan-Meier method and Cox regression.Of 209 patients, 93 (45%) met criteria for DM (35 longstanding DM, 55 new-onset DM, 3 duration unknown). DM patients were older (DM 66 +/- 9 years, non-DM 63 +/- 12 years, P = 0.06); a majority had additional preoperative comorbidities (DM 64.5%, non-DM 25.9%, P0.001). Tumor size was larger in patients with DM (DM 3.8 +/- 1.7 cm, non-DM 3.2 +/- 1.5 cm, P = 0.003). Groups were similar in terms of tumor location, perineural/lymphovascular invasion, and node and margin status. On logistic regression, tumor size/=3.0 cm was independently associated with both overall DM (odds ratio [OR] 3.60; 95% confidence interval [1.79-7.26]) and new-onset DM (OR 3.69, [1.65-8.24]). Median survival was reduced in patients with DM compared with non-DM (15 versus 17 months, P = 0.015). Multivariate analysis controlling for prognostic variables including age, comorbidities, and tumor size demonstrated that DM was independently associated with reduced survival (hazard ratio [HR] 1.55, [1.02-2.35]). This association was more pronounced for patients with new-onset DM (HR 1.75 [1.10-2.78]) than those with longstanding DM (HR 1.30 [0.75-2.25]).Preexisting DM is associated with reduced survival in patients undergoing resection for PDAC. PDAC with new-onset DM may exhibit increased tumor size and decreased postresection survival. Additional investigation is needed to clarify etiology and impact of PDAC-associated DM.

Background The role of adjuvant chemoradiation therapy (CRT) in pancreatic cancer remains controversial. The primary aim of this study was to determine if CRT improved survival in patients with resected pancreatic cancer in a large, multiinstitutional cohort of patients. Study Design Patients undergoing resection for pancreatic adenocarcinoma from seven academic medical institutions were included. Exclusion criteria included patients with T4 or M1 disease, R2 resection margin, preoperative therapy, chemotherapy alone, or if adjuvant therapy status was unknown. Results There were 747 patients included in the initial evaluation. Primary analysis was performed between patients that had surgery alone (n=374) and those receiving adjuvant CRT (n=299). Median followup time was 12.2 months and 14.5 months for survivors. Median overall survival for patients receiving adjuvant CRT was significantly longer than for those undergoing operation alone (20.0 months versus 14.5 months, p=0.001). On subset and multivariate analysis, adjuvant CRT demonstrated a significant survival advantage only among patients who had lymph node (LN)-positive disease (hazard ratio 0.477, 95% CI 0.357 to 0.638) and not for LN-negative patients (hazard ratio 0.810, 95% CI 0.556 to 1.181). Disease-free survival in patients with LN-negative disease who received adjuvant CRT was significantly worse than in patients who had surgery alone (14.5 months versus 18.6 months, p=0.034). Conclusions This large multiinstitutional study emphasizes the importance of analyzing subsets of patients with pancreas adenocarcinoma who have LN metastasis. Benefit of adjuvant CRT is seen only in patients with LN-positive disease, regardless of resection margin status. CRT in patients with LN-negative disease may contribute to reduced disease-free survival.

Modern treatments are prolonging life for metastatic breast cancer patients. Reconstruction in these patients is controversial. The purpose of this study was to characterize de novo metastatic breast cancer patients who undergo mastectomy and reconstruction and to report complication and survival rates. We queried the National Cancer Database for de novo metastatic breast cancer patients, who underwent systemic therapy and mastectomy with reconstruction (R) or without reconstruction (NR) between 2004 and 2013. Patient-tumor characteristics, mortality, and readmissions were compared. Propensity score matched analysis was performed, and survival was calculated using the Kaplan–Meier method. A total of 8554 patients fulfilled study criteria (n = 980/11.5% R vs. n = 7574/88.5% NR). There was a significant increase in reconstruction rates by year: 5.2% in 2004, 14.3% in 2013 (p

Introduction: Breast cancer in women under 40 years of age is rare, accounting for approximately 5% of cases. The disease tends to be more aggressive in younger women. Younger age has been shown to be an independent predictive of breast reconstruction after total mastectomy. Treatment by total mastectomy and reconstruction is examined in relation to patient age. Methods: A retrospective review of all breast cancer patients treated by total mastectomy and reconstruction between 2005 and 2009 was performed by querying a prospective database. Results: A total of 671 patients underwent total mastectomy and reconstruction; of them, 106 (16%) aged

Management of patients with neuroendocrine liver metastasis (NELM) remains controversial. We sought to examine the relative efficacy of surgical management versus intra-arterial therapy (IAT) for NELM and determine factors predictive of survival. A total of 753 patients who had surgery (n = 339) or IAT (n = 414) for NELM from 1985 to 2010 were identified from nine hepatobiliary centers. Clinicopathologic data were assessed with regression modeling and propensity score matching. Most patients had a pancreatic (32%) or a small bowel (27%) primary tumor; 47% had a hormonally active tumor. There were statistically significant differences in characteristics between surgery versus IAT groups (hormonally active tumors: 28 vs. 48%; hepatic tumor burden > 25%: 52% vs. 76%) (all P < 0.001). Among surgical patients, most underwent hepatic resection alone without ablation (78%). The median number of IAT treatments was 1 (range, 1-4). Median and 5-year survival of patients treated with surgery was 123 months and 74% vs. 34 months and 30% for IAT (P < 0.001). In the propensity-adjusted multivariate Cox model, asymptomatic disease (hazard ratio 2.6) was strongly associated with worse outcome (P = 0.001). Although surgical management provided a survival benefit over IAT among symptomatic patients with > 25% hepatic tumor involvement, there was no difference in long-term outcome after surgery versus IAT among asymptomatic patients (P = 0.78). Asymptomatic patients with a large (> 25%) burden of liver disease benefited least from surgical management and IAT may be a more appropriate treatment strategy. Surgical management of NELM should be reserved for patients with low-volume disease or for those patients with symptomatic high-volume disease.

Background and Objectives Biopsy of Cloquet's node (CN) during groin dissection has been used to indicate need for pelvic dissection. With earlier detection of microscopic regional disease in the era of sentinel node biopsy (SNB), frequency of positive CN may be so low that routine biopsy is unwarranted. Methods Patients with positive groin SNB from 2000 to 2008 were identified from two centers. Cases where CN was identified at completion node dissection were selected. Lymphoscintigraphic, surgical, pathologic, and recurrence data were reviewed. Results CN was identified in 53 cases. Median age was 44.5 years (range 7–77); median Breslow depth, 1.98 mm (range 0.5–25.0); % Clark's level IV/V, 90%; and % ulceration, 41.5%. Fifty (94.3%) underwent groin dissection alone; three others underwent concomitant pelvic dissection. Two (3.8%) patients had positive CN; both had additional indications for pelvic dissection. Delayed pelvic recurrence rate was 2/53 (3.8%); both patients had negative CN. In the three patients treated with concurrent groin and pelvic dissection, CN reflected pelvic nodal status in two cases; the third had pelvic metastases despite negative CN. Conclusions After positive SNB, disease involvement of CN is rare. Patients with positive biopsies of CN in the SNB era appear likely to have additional indications for pelvic dissection, minimizing utility of CN biopsy. Routine intraoperative sampling of CN may not be warranted during groin dissection for positive SNB. J. Surg. Oncol. 2010;102:315–320. © 2010 Wiley-Liss, Inc.

Objective Limb salvage in the diabetic population is complex, controversial, costly and variable throughout the world. Within the diabetic population, the indications and morbidity surrounding major leg amputation have not been well defined. Research design and methods A retrospective study of the Georgetown Limb Salvage Registry was conducted to determine outcomes in diabetic patients undergoing proximal forefoot/midfoot (PF/M) amputations and to help define the indications for major leg amputation in ambulatory diabetic patients. Results Of the 937 ambulatory diabetic patients identified during this time period, 808 who presented with superficial ulcers or distal forefoot disease and who did not go on to PF/M or higher amputation were eliminated. Records of 88 patients (92 limbs) in the PF/M amputation group and 25 patients in the below the knee amputation (BKA) group were reviewed (no above knee amputations were performed). At 2 years in the PF/M amputation group, 80% (70/88) of patients were still alive, 73% (68/92) of the limbs remained intact, and 64% of patients (56/88) were ambulatory. In contrast, in the BKA population at 2 years, 52% (13/25) of patients were deceased and 64% (16/25) were ambulating with a prosthetic limb (or had been ambulating at the time of death). The only statistically significant difference between these two groups was the presence of Charcot collapse and rear-foot disease in the BKA group. Conclusion Aggressive efforts at salvage with PF/M amputation procedures should be entertained prior to higher level amputations due to the increased morbidity and mortality evidenced.

Although patients with pancreatic ductal adenocarcinoma (PDAC) frequently require medications to treat pre-existing conditions, the impact of these treatments on outcomes post-resection is unknown. The purpose of this study was to determine the impact of preoperative medications on overall survival after pancreatic resection. Multi-institutional data on preoperative medications and outcomes in patients undergoing resection for PDAC were analyzed. Univariate and multivariate analyses were performed to determine which medications were predictive of early mortality. Of the 518 patients resected for PDAC, 13.3% were being treated preoperatively with insulin, 14.8% were on a statin, 1.7% were on steroids, and 7.6% were on thyroxin. On univariate analysis, patients taking preoperative insulin had a higher 90-day mortality rate relative to those not on insulin (13.0% vs. 4.8%, p = 0.024), and those on a statin had a higher 90-day mortality than those who were not (10.8% vs. 4.6%, p = 0.035). Preoperative steroids and thyroxin were not associated with 90-day mortality (p = 0.409 and p = 0.474, respectively). Insulin and statin use was a stronger predictor of 90-day mortality than history of diabetes (p = 0.101), BMI ≥ 30 (p = 0.166), cardiac disease (p = 0.168), pulmonary disease (p = 1.000), or renal dysfunction (p = 1.000). Older patients also had a higher risk of early postoperative death (p = 0.011). On multivariate analysis, only preoperative insulin usage and statin treatment independently predicted early mortality (odds ratio (OR) = 3.043; 95% confidence interval (CI), 1.256–7.372; p = 0.014, and OR = 2.529; 95% CI, 1.048–6.104; p = 0.039, respectively). Based on the beta coefficients, a simple scoring system was devised to predict survival after resection from preoperative medication use. Zero points were assigned to patients who were on neither insulin nor a statin, one point to those who were on one or the other, and two points to those who were on both insulin and a statin. The score correlated with early postoperative survival (90-day mortality rates of 3.4%, 11.5%, and 13.3% for 0, 1, and 2 points, respectively, p = 0.004). Increasing score was also associated with poorer long-term outcomes, with a median overall survival of 19.6, 15.6, and 11.2 months for 0, 1, and 2 points, respectively (p = 0.002, median follow-up 14.4 months). Patients with PDAC being treated for pre-existing diabetes or hypercholesterolemia with either insulin or statin-based therapy have an increased risk of early postoperative mortality. A simple scoring system based on preoperative medications can be used to predict early and overall survival following resection.

The aim of this study was to examine differences in presentation and outcome in pancreaticoduodenectomy for benign and malignant processes. Pancreaticoduodenectomies performed for pancreatic adenocarcinoma or chronic pancreatitis from 2000 to 2008 were identified from a prospectively maintained database and compared. Pancreaticoduodenectomy revealed adenocarcinoma in 220 patients and benign chronic pancreatitis in 40 patients. Patients with adenocarcinoma were older (64 +/- 10.6 years and 47.6 +/- 10.8 years, respectively, P0.001) and more likely to demonstrate jaundice (92 and 13%, respectively, P0.001), weight loss (76 and 58%, respectively, P = 0.01), and discrete masses (72 and 55%, respectively, P = 0.03). By contrast, chronic pancreatitis presented with increased pain (98 and 49%, respectively, P0.001) and nausea (68 and 32%, respectively, P0.001). Patients with pancreatitis experienced reduced blood loss (227 +/- 156 and 571 +/- 626 mL, respectively, P = 0.05) and transfusion (10 and 42%, respectively, P0.001). Postoperatively, the groups were similar in risk of fistula, gastroparesis, overall morbidity, and mortality. Infection rates were higher in adenocarcinoma (42 and 20%, respectively, P = 0.01). Forty-nine patients underwent pancreaticoduodenectomy for presumptive chronic pancreatitis, of which nine had adenocarcinoma; logistic regression identified increasing age and jaundice as predictors of malignancy. Compared with patients with adenocarcinoma, those who undergo pancreaticoduodenectomy for chronic pancreatitis are less likely to require preoperative biliary drainage and perioperative transfusion. Infectious complication risk is higher in patients with adenocarcinoma. Increased age and jaundice should raise suspicion of cancer in patients with chronic pancreatitis undergoing evaluation for surgical treatment.

The premise that allogeneic red blood cell transfusion (RBCT) contributes to adverse oncologic outcomes after surgery remains controversial. We examined the effects of RBCT during and after pancreaticoduodenectomy (PD) for pancreatic ductal adenocarcinoma (PDAC) on disease recurrence and survival. A prospective database of 220 patients undergoing PD for PDAC from 2000 to 2008 was reviewed and transfusion data collected. Univariate and multivariate analyses were performed for factors influencing RBCT, recurrence-free survival (RFS), and overall survival (OS). The effect of amount and timing (intraoperative vs. postoperative) of RBCT was analyzed. One hundred forty-seven patients (67%) received RBCT: 70 (32%) received 1 to 2 units, and 77 (35%) received >2 units. Median RFS and OS for the entire cohort was 12 and 16 months, respectively. RBCT of >2 units was associated with reduced RFS (9 vs. 15 months; P = 0.033) and OS (14 vs. 20 months; P = 0.003). Stratified by timing of transfusion, postoperative RBCT was associated with shortened RFS and OS. Controlling for age, body mass index, comorbidities, tumor factors, and major complications, each incremental unit of postoperative RBCT was associated with reduced RFS (hazard ratio 1.10, 95% confidence interval 1.02–1.18) and OS (hazard ratio 1.08, 95% confidence interval 1.03–1.12). Low hemoglobin and presence of comorbidities were the only preoperative factors independently associated with RBCT. Allogeneic red blood cell transfusion after PD for PDAC is independently associated with earlier cancer recurrence and reduced survival, in particular when administered postoperatively and in larger quantities. Blood-conservation methods are especially indicated for patients with preoperative anemia and medical comorbidities.

Background Patients with pancreatic cancer who present with biliary symptoms may undergo cholecystectomy and thus delay cancer diagnosis. We hypothesized that prior cholecystectomy leads to decreased overall survival in patients with pancreatic adenocarcinoma. Methods Retrospective study of hepatobiliary database. Results Three hundred sixty-five patients with a diagnosis of resectable periampullary pancreatic adenocarcinoma were identified. Eighty-seven patients underwent prior cholecystectomy. Median age ( P = .48), body mass index (BMI) ( P = .8), diabetes status ( P = .06), American Society of Anesthesiologists (ASA) class ( P = .22), stent placement ( P = .13), operative time ( P = .76), estimated blood loss (EBL) ( P = .24), intraoperative transfusion ( P = .91), portal vein resection ( P = .25), LOS ( P = .09) adjuvant therapy ( P = .2), tumor size ( P = .89), differentiation ( P = .67), angiolymphatic invasion ( P = .69), perineural invasion ( P = 54), nodal metastasis ( P = .43), complication rate ( P = .75), and 30-day mortality ( P = .58) were not statistically different between patients with previous cholecystectomy and those without. Median survival was 14 months for patients with a history of cholecystectomy and 16 months for those without ( P = .25). Previous cholecystectomy was not a predictor of survival on Cox regression analysis. Conclusion There was no difference in overall survival in patients with pancreatic cancer with prior cholecystectomy versus those without.

Background Associations between diabetes mellitus (DM) and pancreatic ductal adenocarcinoma (PDAC) are well established; however, the impact of DM on perioperative morbidity and mortality after PDAC resection is unclear. Study Design A prospectively maintained database review identified all patients undergoing PDAC resection between January 1, 2000, and December 31, 2008. Diagnosis of DM was established by history and biochemical profile. Records were reviewed for comorbidities, operative variables, and histologic parameters. Postoperative morbidity and mortality were assessed for diabetic and nondiabetic patients using standardized definitions. Results Of 251 PDAC cases, 116 (46%) patients had preoperative DM. Pancreaticoduodenectomy was performed in 220 (87.6%), left pancreatectomy in 29 (11.6%), and total pancreatectomy in 2 (0.8%). The major complication rate was 25.5%, with 60-day mortality of 3.6%. Delayed gastric emptying (DGE) occurred in 40.1% of patients. Pancreatic fistulas developed in 17 (6.8%) patients; 11 of them were clinically significant (grades B/C). DM patients had a higher likelihood of developing fistulas (DM 10.3%, non-DM 3.7%, p=0.04). When controlled for age, comorbidities, body mass index, preoperative albumin level, operation type, operative time, and pancreatic quality, DM maintained an independent association with fistula formation (odds ratio 4.3, 95% CI 1.18 to 15.8, p=0.027). Acute kidney injury was more frequent in the DM group (DM 23.3%, non-DM 12.6%, p=0.03). DM and non-DM patients had similar frequency of DGE, wound infections, intra-abdominal abscesses, and cardiovascular and pulmonary complications, as well as length of stay and mortality. Conclusions Comorbid DM does not influence perioperative outcomes dramatically after pancreatectomy for ductal adenocarcinoma. The role of PDAC-associated DM as a risk factor for postresection pancreatic fistula should be further explored. Evaluation of glycemic control and outcomes after PDAC resection may be useful.

Ampullary adenocarcinoma (AmpCA) carries a better overall survival (OS) rate than other periampullary cancers. We examined clinicopathologic features in AmpCA for impact on OS. Records of patients undergoing pancreaticoduodenectomy from 2000 to 2007 for AmpCA were reviewed and histological specimens were reanalyzed. Of 302 patients undergoing pancreaticoduodenectomy for malignancy, 45 (14.9%) had AmpCA. Mean age was 61.3 +/- 12.2 years, mean tumor size was 2.6 +/- 1.3 cm, 57 per cent wereor = T3 tumors, 42 per cent were N1 stage, 13 (49%) had perineural invasion (PNI), and 29 (64%) had lymphovascular invasion (LVI). Thirteen were intestinal (29%), 14 were pancreaticobiliary (31%), and 18 were mixed (40%). Median OS was 42 months (range 4-80 mos). On log rank testing,or = T3 (24 vs 65 mos, P0.01), N1 (25 vs 61 mos, P0.01), poor differentiation (24 vs 44 mos, P = 0.01), pancreaticobiliary subtype (23 vs 44 mos, P = 0.01), and PNI (23 vs 44 mos, P0.01) were significant for worse survival. By multivariate analysis, N1 disease (hazard ratio [HR] 4.50, 95% confidence interval [CI] 1.16-17.40) and PNI (HR 4.62, CI 1.11-19.21) maintained associations with worse survival, whereas histological subtype did not. N1 disease and presence of PNI demonstrated independent associations with worse survival. Given high percentage of mixed histology, PNI may be more informative than the subtype in predicting outcome for patients with AmpCA.