Inter-observer and intra-observer reliability between manual segmentation and semi-automated segmentation for carotid vessel wall volume measurements on three-dimensional ultrasonography.

PURPOSE: Carotid vessel wall volume (VWV) measurement on three-dimensional ultrasonography (3DUS) outperforms conventional two-dimensional ultrasonography for carotid atherosclerosis evaluation. Although time-saving semi-automated algorithms have been introduced, their clinical availability remains limited due to a lack of validation, particularly an extensive reliability analysis. This study compared inter-observer and intra-observer reliability between manual segmentation and semi-automated segmentation for carotid VWV measurements on 3DUS. METHODS: Thirty-one 3DUS volume datasets were prospectively acquired from 20 healthy subjects, aged >18 years, without previous stroke, transient ischemic attack, or cardiovascular disease. Five observers segmented all volume datasets both manually and semi-automatically. The process was repeated five times. Reliability was expressed by the intraclass correlation coefficient, supplemented by the coefficient of variation. RESULTS: Carotid VWV measurements using the common carotid artery (CCA) were more reliable than those using the internal carotid artery (ICA) or external carotid artery (ECA) for both manual and semiautomated segmentation (manual segmentation, CCA: inter-observer, 0.935; intra-observer, 0.934 to 0.966; ICA: inter-observer, 0.784; intra-observer, 0.756 to 0.878; ECA: inter-observer, 0.732; intraobserver, 0.919 to 0.962; semi-automated segmentation, CCA: inter-observer, 0.986; intra-observer, 0.954 to 0.993; ICA: inter-observer, 0.977; intra-observer, 0.958 to 0.978; ECA: inter-observer, 0.966; intra-observer, 0.884 to 0.937). Total carotid VWV measurements by manual (inter-observer, 0.922; intra-observer, 0.927 to 0.961) and semi-automated segmentation (inter-observer, 0.987; intra-observer, 0.968 to 0.989) were highly reliable. Semi-automated segmentation showed higher reliability than manual segmentation for both individual and total carotid VWV measurements. CONCLUSION: 3DUS carotid VWV measurements of the CCA are more reliable than measurements of the ICA and ECA. Total carotid VWV measurements are highly reliable. Semi-automated segmentation has higher reliability than manual segmentation.

Robotic Rectus Abdominis Harvest for Pelvic Reconstruction after Abdominoperineal Resection.

The use of robotic surgical systems to perform abdominoperineal resection (APR) has recently become more prevalent. This minimally invasive approach produces fewer scars and potentially less morbidity for the patient. The rectus abdominis muscle is often used for reconstruction after APR if primary closure is not feasible or the surgical site is at high risk of wound complications. Since the traditional open harvest of this flap creates large incisions that negate the advantages of minimally invasive APR, there has been growing interest in harvesting the rectus abdominis in a similarly robotic fashion. This article reviews the technique, benefits, and limitations of this robotic technique. Compared to the traditional open harvest, robotic harvest of the rectus abdominis leaves smaller scars, provides technical benefits for the surgeon, and offers possible morbidity benefits for the patient. These advantages should be weighed against the added expense and learning curve inherent to robotic surgery.

Factors associated with inability to return to intended oncologic treatment in pancreatic cancer.

BACKGROUND: Return to Intended Oncologic Treatment (RIOT) has been proposed as a quality metric in the care of cancer patients. We sought to define factors associated with inability to RIOT in Pancreatic Ductal Adenocarcinoma (PDAC) patients. METHODS: The NCDB was queried for patients who underwent pancreaticoduodenectomy for pathologic stage IB, IIA, or IIB PDAC from 2010 to 2016. Multivariable binary logistic regression models identified factors associated with failure to RIOT, and Kaplan-Meier survival analysis and Cox multivariable regression models demonstrated the impact of failure to RIOT on survival. RESULTS: Increasing age (p < .001), Hispanic race (p = .002), pathological stage IB (p = .004) and IIA (p = .001) as compared to IIB, increasing hospital stay (p < .001), and open surgical approach (p = .024) were associated with increased risk of inability to RIOT. Male sex (p < .001), Charlson-Deyo scores of 0 (p < .001) and 1 (p = .001) as compared to >2, negative surgical margins (p = .048), receiving care at academic institutions (p = .001), and increasing institutional case volume (p = .001) were associated with improved odds of RIOT. CONCLUSIONS: Patient features can impact RIOT and should be considered when designing multi-modality treatment strategies.

Malnutrition in elective surgery: How traditional markers might be failing surgeons and patients.

BACKGROUND: Serologic and anthropometric measures are commonly used as surrogate markers of nutritional status in clinical practice. In 2012, leading dietetic organizations published a standard definition of malnutrition based on clinical characteristics. We hypothesize that surrogate markers underrecognize clinical malnutrition and do not accurately identify patients at risk for adverse outcomes. METHODS: A single-institution cohort study of elective surgical inpatients from August 2015 to November 2017. Nutritional assessment was completed by trained registered dietitians using leading dietetic guidelines. Multivariable logistic regression was used to determine the association between malnutrition and perioperative outcomes. RESULTS: Among 953 elective surgical admissions, 456 underwent full clinical nutritional assessment. Of these, 202 (44.3%) met malnutrition criteria. In addition, 20.3% of patients with clinical malnutrition were underweight (<18.5 kg/m2) and 38.1% had a serum albumin <3.0 g/dL. Compared with nonmalnourished patients, those with clinical malnutrition had higher rates of any complication (46.5% vs 37.8%, P = .06), overall infectious complications (26.2% vs 14.6%, P = .002), surgical site infections (9.4% vs 3.9%, P = .02), and mortality (8.9% vs 1.9%, P = .001). Clinical malnutrition was associated with death (odds ratio 3.99; 95% confidence interval, 1.27-12.54), overall infectious complication (odds ratio 1.77; 95% confidence interval, 1.07-2.94), and surgical site infections (odds ratio 2.65; 95% confidence interval, 1.12-6.22). CONCLUSION: In this cohort of elective surgical patients, traditional markers failed to identify malnutrition in a substantial portion of patients who met clinical malnutrition criteria. Clinical malnutrition assessment is effective in identifying patients who may be at risk for suboptimal outcomes. Surgeons should implement clinical nutritional assessment and factor that information into their preoperative evaluation and management of elective surgical patients.

Variation in surgical management of primary hyperparathyroidism in the US Department of Veterans Affairs healthcare system: A 15-year observational study.

BACKGROUND: Primary hyperparathyroidism is underdiagnosed and undertreated nationally despite the benefits of parathyroidectomy. However, the degree of hospital-level variation in the management of primary hyperparathyroidism is unknown. METHODS: We performed a national, retrospective study of Veterans with primary hyperparathyroidism using the Veterans Affairs Corporate Data Warehouse from January 2000 to September 2015. The objective was to characterize the extent of hospital-level variation in the use of parathyroidectomy for the management of primary hyperparathyroidism within a national, integrated healthcare system. Rate of parathyroidectomy in patients with primary hyperparathyroidism was stratified by (1) geographic region, (2) facility complexity level, (3) volume of parathyroidectomies per facility, and (4) frequency of parathyroid hormone testing in hypercalcemic patients. RESULTS: Among 47,158 Veterans with primary hyperparathyroidism, 6,048 (12.8%) underwent parathyroidectomy. Rates of parathyroidectomy were significantly higher in the Continental (17.0%) and Pacific (16.0%) regions than in other areas (11.4%, P < .01). The highest complexity referral centers had the highest rate of parathyroidectomy (13.6%) compared with all other facilities (12.1%, P < .01). Centers that performed the highest volume of parathyroidectomies were more likely to offer surgery (13.3%) than low volume centers (8.9%, P < .01). Facilities with higher frequency of parathyroid hormone testing among hypercalcemic patients were more likely to offer parathyroidectomy (15.2%) than those with the lowest parathyroid hormone testing frequency (12.6%, P < .01). CONCLUSION: Although there is notable variation in parathyroidectomy use for definitive treatment of primary hyperparathyroidism between Veterans Affairs facilities, parathyroidectomy rates are low across the entire system. Further research is needed to understand additional local contextual and other patient and clinician-level factors for the undertreatment of primary hyperparathyroidism to subsequently guide corrective interventions.

Nodal downstaging as a treatment goal for node-positive pancreatic cancer.

BACKGROUND: Nodal metastases portend a poor prognosis in patients with localized pancreatic cancer. Neoadjuvant therapy is associated with pathologic nodal downstaging in up to 38% of patients. However, the optimal type of neoadjuvant therapy for achieving nodal downstaging is unclear. METHODS: We conducted a retrospective cohort study of patients with nonmetastatic, clinically node-positive pancreatic cancer treated with neoadjuvant therapy and surgery identified in the National Cancer Database (2006-2014). Patients were stratified based on the neoadjuvant therapy regimens they received: multiagent chemotherapy; single-agent chemotherapy; multiagent chemotherapy with radiation; and single-agent chemotherapy with radiation. Associations between nodal downstaging and the type of neoadjuvant therapy received and overall risk of death were evaluated using multivariable regression analyses. RESULTS: Among the 603 pancreatic ductal adenocarcinoma patients treated with neoadjuvant therapy, 400 received multiagent chemotherapy (202 with radiation) and 203 received single agent chemotherapy (151 with radiation). Relative to multiagent chemotherapy, single-agent chemotherapy was associated with a lower likelihood of nodal downstaging (relative risk ratio 0.38 [95% CI 0.17-0.85]). Use of radiation was associated with a significantly greater likelihood of nodal response (single-agent chemotherapy with radiation: relative risk ratio 1.77 [1.36-2.30]; multiagent chemotherapy with radiation: relative risk ratio 1.91 [1.49-2.45]; radiation use overall (versus no radiation): relative risk ratio 2.12 [1.68-2.68]). Compared with patients who remained pathologically node positive after neoadjuvant therapy, node negative status was associated with a significantly lower risk of death (hazard ratio 0.61 [0.49-0.76]) regardless of whether radiation was used (hazard ratio 0.63 [0.48-0.82]) or not (hazard ratio 0.45 [0.29-0.72]). CONCLUSION: Nodal downstaging is associated with a survival benefit in patients with node-positive pancreatic ductal adenocarcinoma and is most likely to be achieved with neoadjuvant therapy that includes radiation. Single-agent chemotherapy neoadjuvant therapy was least likely to result in nodal downstaging.

Understaging of clinical stage I pancreatic cancer and the impact of multimodality therapy.

BACKGROUND: Although current guidelines recommend multimodal therapy for all patients with pancreatic ductal adenocarcinoma, it is unclear the extent to which clinical stage I patients are accurately staged and how this may affect management. METHODS: In this retrospective cohort study of 4,404 patients aged 18-79 years with clinical stage 1 (ie, T1N0 or T2N0) pancreatic ductal adenocarcinoma treated with upfront resection in the National Cancer Database (2004-2014), understaging was ascertained by comparing pretreatment clinical stage with pathologic stage. The association between adjuvant treatment and overall risk of death among true stage I and understaged patients was evaluated using multivariable Cox regression. RESULTS: Upstaging was identified in 72.6% of patients (62.8% T3/4, 53.9% N1) of whom 69.7% received adjuvant therapy compared with 47.0% with true stage I disease. Overall survival at 5 years among those with true stage I disease was significantly higher than those who had been clinically understaged (42.9% vs 16.6%; log-rank, p < 0.001). For true stage I patients, adjuvant therapy was not associated with risk of death (hazard ratio: 1.07, 95% confidence interval: 0.89-1.29). For understaged patients, adjuvant therapy significantly decreased risk of death (hazard ratio: 0.64, 95% confidence interval: 0.55-0.74). CONCLUSION: The majority of clinical stage I pancreatic ductal adenocarcinoma patients actually have higher-stage disease and benefit from multimodal therapy; however, one third of understaged patients do not receive any adjuvant treatment. Clinicians should discuss all potential treatment strategies with patients (in the context of the acknowledged risks and benefits), including the utilization of neoadjuvant approaches in those presenting with potentially resectable disease.

National Failure of Surgical Staging for T1b Gallbladder Cancer.

BACKGROUND: Current guidelines recommend radical cholecystectomy with regional lymphadenectomy (RC-RL) for patients with T1b gallbladder cancer (GBC). However, the extent to which these guidelines are followed is unclear. This study aimed to evaluate current surgical practices for T1b GBC and their implications for overall management strategies and associated outcomes. METHODS: This retrospective cohort study investigated patients identified from the National Cancer Data Base (2004-2012) with non-metastatic T1b GBC. The patients were categorized according to type of surgical treatment received: simple cholecystectomy (SC) or RC-RL. Among the patients who had lymph nodes pathologically examined, nodal status was classified as pN- or pN+. Use of any adjuvant therapy was ascertained. Overall survival (OS) was compared based on type of surgical treatment and nodal status. RESULTS: The cohort comprised 464 patients (247 SC and 217 RC-RL cases). The positive margin status did not differ between the two groups (6.1% for SC vs 2.3% for RC-RL; p = 0.128). For RC-RL, the pN+ rate was 15%. Adjuvant therapies were used more frequently in pN+ (53.1% vs 9.4% for pN-). By comparison, 10.9% of the SC patients received adjuvant therapy. The OS for RC-RL-pN- (5-years OS, 64.4%) was significantly better than for RC-RL-pN+ (5-years OS, 15.7%) or SC (5-years OS, 48.3%) (p < 0.001). CONCLUSION: Less than 50% of the patients with a T1b GBC primary tumor undergo the recommended surgical treatment. Given that 15% of these patients have nodal metastasis and in light of the previously described benefits of adjuvant therapy for node positive GBC, failure to perform RC-RL risks incomplete staging and thus undertreatment for patients with T1b GBC.

Non-operative management of iatrogenic colonic perforation after percutaneous cholecystotomy.

The management of iatrogenic colonic perforation encountered during percutaneous cholecystotomy tube placement is not well reported. It is unclear as to whether an operative versus a conservative approach is ideal for this complication. We therefore present our case report to spur a discussion on patient selection, interval follow-up and call for future studies regarding this uncommon complication.

Robotic port-site hernias after general surgical procedures.

BACKGROUND: With the increasing use of the robotic platform in general surgery, whether 8-mm ports should be closed comes into question. We sought to characterize the incidence of port-site hernias (PSHs) among patients undergoing robotic-assisted general surgery. METHODS: A retrospective chart review of a single institutional database identified patients who underwent robotic-assisted general surgery from July 2010 to December 2016. For each patient, the number, type, location, and size of all ports were collected. Twelve-millimeter port sites were routinely closed, whereas 5-mm and 8-mm port sites were not. PSH was detected on review of documented physical examination and of postoperative cross-sectional imaging, when available, in which case it was defined as a disruption of the fascia with or without eventration of tissue at a site of prior port placement. RESULTS: One hundred and seventy-eight patients underwent robotic-assisted general surgery, with 725 total ports: 433 8-mm working ports, 72 12-mm working ports, 178 12-mm camera ports, and 42 5-mm assistant ports. Ninety-four percent of the patients were men, the mean age was 63 ± 12, body mass index was 29 ± 7 kg/m2, and the median American Society of Anesthesiologists score was 3. Types of cases included 68 rectal (38.2%), 36 colon (20.2%), 25 hepatopancreatobiliary (14.0%), 21 inguinal hernia (11.8%), and 28 "other" (15.7%) operations. At a median follow-up of 193 d, there were three PSHs through 8-mm port sites (0.7%), two PSHs through 12-mm port sites (0.8%), and no PSH through 5-mm port sites. Two of the three 8-mm PSHs occurred in the early postoperative period and required emergent repair due to small bowel incarceration. CONCLUSIONS: PSHs through 8-mm robotic port sites occur infrequently but can cause significant morbidity. Further investigation with longer follow-up is warranted to better understand the true incidence of robotic PSH.

The role of surgery and adjuvant therapy in lymph node-positive cancers of the gallbladder and intrahepatic bile ducts.

BACKGROUND: Lymph node metastasis is a poor prognostic factor for biliary tract cancers (BTCs). The optimal management of patients who have BTC with positive regional lymph nodes, including the impact of surgery and adjuvant therapy (AT), is unclear. METHODS: This was a retrospective cohort study of patients who had T1-T3N1M0 gallbladder cancer (GBC) and intrahepatic cholangiocarcinoma (IHC) in the National Cancer Database (2004-2012). Patients were classified by treatment approach (nonoperative, surgery, surgery plus AT). Associations between the overall risk of death and treatment strategy were evaluated with multivariable Cox regression. RESULTS: Rates of surgical resection were 84.1% for patients with GBC (n = 1335) and 36.6% for those with IHC (n = 1009). The median overall survival of patients in the nonoperative, surgery, and surgery plus AT group was 11.6, 13.3, and 19.6 months, respectively, for those with GBC (log-rank P < .001), and 12.7, 16.2, and 22.6 months, respectively, for those with IHC (log-rank P < .001), respectively. Compared with nonoperative therapy, surgery with or without AT was associated with a lower risk of death from GBC (surgery with AT: hazard ratio [HR], 0.59; 95% confidence interval [CI], 0.48-0.73; surgery without AT: HR, 0.71; 95% CI, 0.56-0.89) and from IHC (surgery with AT: HR, 0.52; 95% CI, 0.42-0.63; surgery without AT: HR, 0.70; 95% CI, 0.56-0.87). AT that included radiation was associated with a lower risk of death relative to surgery alone for patients with GBC regardless of margin status (margin-negative resection: HR, 0.66; 95% CI, 0.52-0.84; margin-positive resection: HR, 0.54; 95% CI, 0.39-0.75), but adjuvant chemotherapy alone was not. For patients with IHC, no survival benefit was detected with adjuvant chemotherapy or radiation for those who underwent either margin-positive or margin-negative resection. CONCLUSIONS: The best outcomes for patients who have lymph node-positive BTCs are associated with margin-negative resection and, in those who have GBC, the inclusion of adjuvant chemotherapy with radiation regardless of margin status. Cancer 2018;124:74-83. © 2017 American Cancer Society.

Association of the Addition of Oral Antibiotics to Mechanical Bowel Preparation for Left Colon and Rectal Cancer Resections With Reduction of Surgical Site Infections.

IMPORTANCE: Surgical site infections (SSIs) after colorectal surgery remain a significant complication, particularly for patients with cancer, because they can delay the administration of adjuvant therapy. A combination of oral antibiotics and mechanical bowel preparation (MBP) is a potential, yet controversial, SSI prevention strategy. OBJECTIVE: To determine the association of the addition of oral antibiotics to MBP with preventing SSIs in left colon and rectal cancer resections and its association with the timely administration of adjuvant therapy. DESIGN, SETTING, AND PARTICIPANTS: A retrospective review was performed of 89 patients undergoing left colon and rectal cancer resections from October 1, 2013, to December 31, 2016, at a single institution. A bowel regimen of oral antibiotics and MBP (neomycin sulfate, metronidazole hydrochloride, and magnesium citrate) was implemented August 1, 2015. Patients receiving MBP and oral antibiotics and those undergoing MBP without oral antibiotics were compared using univariate analysis. Multivariable logistic regression controlling for factors that may affect SSIs was used to evaluate the association between use of oral antibiotics and MBP and the occurrence of SSIs. MAIN OUTCOMES AND MEASURES: Surgical site infections within 30 days of the index procedure and time to adjuvant therapy. RESULTS: Of the 89 patients (5 women and 84 men; mean [SD] age, 65.3 [9.2] years) in the study, 49 underwent surgery with MBP but without oral antibiotics and 40 underwent surgery with MBP and oral antibiotics. The patients who received oral antibiotics and MBP were younger than those who received only MBP (mean [SD] age, 62.6 [9.1] vs 67.5 [8.8] years; P = .01), but these 2 cohorts of patients were otherwise similar in baseline demographic, clinical, and cancer characteristics. Surgical approach (minimally invasive vs open) and case type were similarly distributed; however, the median operative time of patients who received oral antibiotics and MBP was longer than that of patients who received MBP only (391 minutes [interquartile range, 302-550 minutes] vs 348 minutes [interquartile range, 248-425 minutes]; P = .03). The overall SSI rate was lower for patients who received oral antibiotics and MBP than for patients who received MBP only (3 [8%] vs 13 [27%]; P = .03), with no deep or organ space SSIs or anastomotic leaks in patients who received oral antibiotics and MBP compared with 9 organ space SSIs (18%; P = .004) and 5 anastomotic leaks (10%; P = .06) in patients who received MBP only. Despite this finding, there was no difference in median days to adjuvant therapy between the 2 cohorts (60 days [interquartile range, 46-73 days] for patients who received MBP only vs 72 days [interquartile range, 59-85 days] for patients who received oral antibiotics and MBP; P = .13). Oral antibiotics and MBP (odds ratio, 0.11; 95% CI, 0.02-0.86; P = .04) and minimally invasive surgery (odds ratio, 0.22; 95% CI, 0.05-0.89; P = .03) were independently associated with reduced odds of SSIs. CONCLUSIONS AND RELEVANCE: The combination of oral antibiotics and MBP is associated with a significant decrease in the rate of SSIs and should be considered for patients undergoing elective left colon and rectal cancer resections.

A Prospective Randomized Multicenter Trial of Distal Pancreatectomy With and Without Routine Intraperitoneal Drainage.

OBJECTIVE: The objective of this study was to test the hypothesis that distal pancreatectomy (DP) without intraperitoneal drainage does not affect the frequency of grade 2 or higher grade complications. BACKGROUND: The use of routine intraperitoneal drains during DP is controversial. Prior to this study, no prospective trial focusing on DP without intraperitoneal drainage has been reported. METHODS: Patients undergoing DP for all causes at 14 high-volume pancreas centers were preoperatively randomized to placement of a drain or no drain. Complications and their severity were tracked for 60 days and mortality for 90 days. The study was powered to detect a 15% positive or negative difference in the rate of grade 2 or higher grade complications. All data were collected prospectively and source documents were reviewed at the coordinating center to confirm completeness and accuracy. RESULTS: A total of 344 patients underwent DP with (N = 174) and without (N = 170) the use of intraperitoneal drainage. There were no differences between cohorts in demographics, comorbidities, pathology, pancreatic duct size, pancreas texture, or operative technique. There was no difference in the rate of grade 2 or higher grade complications (44% vs. 42%, P = 0.80). There was no difference in clinically relevant postoperative pancreatic fistula (18% vs 12%, P = 0.11) or mortality (0% vs 1%, P = 0.24). DP without routine intraperitoneal drainage was associated with a higher incidence of intra-abdominal fluid collection (9% vs 22%, P = 0.0004). There was no difference in the frequency of postoperative imaging, percutaneous drain placement, reoperation, readmission, or quality of life scores. CONCLUSIONS: This prospective randomized multicenter trial provides evidence that clinical outcomes are comparable in DP with or without intraperitoneal drainage.

Value of lymph node positivity in treatment planning for early stage pancreatic cancer.

BACKGROUND: Multimodal therapy is recommended for early stage pancreatic cancer, although whether all patients benefit and the optimal timing of chemotherapy remain unclear. METHODS: Retrospective cohort study of patients aged 18-79 years with stage I-II pancreatic ductal adenocarcinoma in the National Cancer Database (2004-2012). Patients were grouped based on treatment strategy as surgery only, adjuvant, and preoperative. Accuracy of nodal staging and rate of nodal downstaging were ascertained using pretreatment clinical and postresection pathologic nodal status data. Association between overall risk of death and treatment strategy was evaluated with multivariable Cox regression. RESULTS: Among 19,031 patients, 31.1% underwent surgery only, 59.6% received adjuvant, and 9.3% preoperative therapy. Based on patients receiving upfront surgery, clinical nodal staging bore sensitivity, specificity, positive predictive value, and negative predictive value of 46.2%, 95.7%, 95.1%, and 49.8%, respectively. Preoperative therapy downstaged 38% of cN1 patients to ypN0; 5-year overall survival for this group was 27.2% vs 12.3% for ypN1 patients (P < .001). Relative to surgery only, adjuvant (HR 0.75, 95% CI [0.71-0.78]) and preoperative therapy (HR 0.66 [0.60-0.73]) were associated with lower risk of death among patients with pN1, but not pN0 (adjuvant-HR 1.01 [0.94-1.09]; preoperative-HR 1.10 [0.99-1.22]), disease. CONCLUSION: Our data provide strong support for preoperative chemotherapy for patients with node-positive pancreatic cancer, one third of whom may be downstaged. Among those with seemingly node-negative disease, half will be understaged with current clinical staging modalities. These results should be considered when planning treatment for patients with early stage pancreatic cancer.

National Quality Forum Colon Cancer Quality Metric Performance: How Are Hospitals Measuring Up?

OBJECTIVE: To evaluate the impact of care at high-performing hospitals on the National Quality Forum (NQF) colon cancer metrics. BACKGROUND: The NQF endorses evaluating ≥12 lymph nodes (LNs), adjuvant chemotherapy (AC) for stage III patients, and AC within 4 months of diagnosis as colon cancer quality indicators. Data on hospital-level metric performance and the association with survival are unclear. METHODS: Retrospective cohort study of 218,186 patients with resected stage I to III colon cancer in the National Cancer Data Base (2004-2012). High-performing hospitals (>75% achievement) were identified by the proportion of patients achieving each measure. The association between hospital performance and survival was evaluated using Cox shared frailty modeling. RESULTS: Only hospital LN performance improved (15.8% in 2004 vs 80.7% in 2012; trend test, P < 0.001), with 45.9% of hospitals performing well on all 3 measures concurrently in the most recent study year. Overall, 5-year survival was 75.0%, 72.3%, 72.5%, and 69.5% for those treated at hospitals with high performance on 3, 2, 1, and 0 metrics, respectively (log-rank, P < 0.001). Care at hospitals with high metric performance was associated with lower risk of death in a dose-response fashion [0 metrics, reference; 1, hazard ratio (HR) 0.96 (0.89-1.03); 2, HR 0.92 (0.87-0.98); 3, HR 0.85 (0.80-0.90); 2 vs 1, HR 0.96 (0.91-1.01); 3 vs 1, HR 0.89 (0.84-0.93); 3 vs 2, HR 0.95 (0.89-0.95)]. Performance on metrics in combination was associated with lower risk of death [LN + AC, HR 0.86 (0.78-0.95); AC + timely AC, HR 0.92 (0.87-0.98); LN + AC + timely AC, HR 0.85 (0.80-0.90)], whereas individual measures were not [LN, HR 0.95 (0.88-1.04); AC, HR 0.95 (0.87-1.05)]. CONCLUSIONS: Less than half of hospitals perform well on these NQF colon cancer metrics concurrently, and high performance on individual measures is not associated with improved survival. Quality improvement efforts should shift focus from individual measures to defining composite measures encompassing the overall multimodal care pathway and capturing successful transitions from one care modality to another.