Predictors of residual disease after unplanned excision of soft tissue sarcomas.

BACKGROUND: Unplanned excision of soft tissue sarcomas (STS) is an important quality of care issue given the morbidity related to tumor bed excision. Since not all patients harbor residual disease at the time of reexcision, we sought to determine predictors of residual STS following unplanned excision. METHODS: We identified 76 patients from a prospective database (January 1, 2008-September 30, 2014) who received a diagnosis of primary STS following unplanned excision on the trunk or extremities. We used univariable and multivariable analyses to evaluate predictors of residual STS as the primary endpoint. We calculated the sensitivity, specificity, and accuracy of interval magnetic resonance imaging (MRI) to predict residual sarcoma at reexcision. RESULTS: Mean age was 52 y, and 63.2% were male. 50% had fragmented unplanned excision. Among patients undergoing reexcision, residual STS was identified in 70%. On univariable analysis, MRI showing gross disease and fragmented excision were significant predictors of residual STS (odds ratio, 10.59; 95% CI, 2.14-52.49; P = 0.004 and odds ratio, 3.61; 95% CI, 1.09-11.94; P = 0.035, respectively). On multivariable analysis, tumor size predicted distant recurrence and overall survival. When we combined equivocal and positive MRI, the sensitivity and specificity of MRI for predicting residual STS were 86.7% (95% CI, 73.2%-95.0%) and 57.9% (95% CI, 33.5%-79.8%), with an overall accuracy of 78.1% (95% CI, 66.0%-87.5%). CONCLUSIONS: About 70% of patients undergoing repeat excision after unplanned excision of STS harbor residual sarcoma. Although interval MRI and fragmented excision appear to be the most significant predictors of residual STS, the accuracy of MRI remains modest, especially given the incidence of equivocal MRI.

Relative value units poorly correlate with measures of surgical effort and complexity.

BACKGROUND: The relationship between procedural relative value units (RVUs) for surgical procedures and other measures of surgeon effort are poorly characterized. We hypothesized that RVUs would poorly correlate with quantifiable metrics of surgeon effort. METHODS: Using the 2010 American College of Surgeons - National Surgical Quality Improvement Program (NSQIP) database, we selected 11 primary current procedural terminology codes associated with high volume surgical procedures. We then identified all patients with a single reported procedural RVU who underwent nonemergent, inpatient general surgical operations. We used linear regression to correlate length of stay (LOS), operative time, overall morbidity, frequency of serious adverse events (SAEs), and mortality with RVUs. We used multivariable logistic regression using all preoperative NSQIP variables to determine other significant predictors of our outcome measures. RESULTS: Among 14,481 patients, RVUs poorly correlated with individual LOS (R(2) = 0.05), operative time (R(2) = 0.10), and mortality (R(2) = 0.35). There was a moderate correlation between RVUs and SAEs (R(2) = 0.79) and RVUs and overall morbidity (R(2) = 0.75). However, among low- to mid-level RVU procedures (11-35) there was a poor correlation between SAEs (R(2) = 0.15), overall morbidity (R(2) = 0.05), and RVUs. On multivariable analysis, RVUs were significant predictors of operative time, LOS, and SAEs (odds ratio 1.06, 95% confidence interval: 1.05-1.07), but RVUs were not a significant predictor of mortality (odds ratio 1.02, 95% confidence interval: 0.99-1.05). CONCLUSIONS: For common, index general surgery procedures, the current RVU assignments poorly correlate with certain metrics of surgeon work, while moderately correlating with others. Given the increasing emphasis on measuring and tracking surgeon productivity, more objective measures of surgeon work and productivity should be developed.

Text messaging among residents and faculty in a university general surgery residency program: prevalence, purpose, and patient care.

BACKGROUND: There is little information about the use of text messaging (texting) devices among resident and faculty physicians for patient-related care (PRC). OBJECTIVE: To determine the prevalence, frequency, purpose, and concerns regarding texting among resident and attending surgeons and to identify factors associated with PRC texting. DESIGN: E-mail survey. SETTING: University medical center and its affiliated hospitals. PARTICIPANTS: Surgery resident and attending staff. OUTCOME MEASURES: Prevalence, frequency, purpose, and concerns regarding patient-related care text messaging. RESULTS: Overall, 73 (65%) surveyed physicians responded, including 45 resident (66%) and 28 attending surgeons (62%). All respondents owned a texting device. Majority of surgery residents (88%) and attendings (71%) texted residents, whereas only 59% of residents and 65% of attendings texted other faculty. Most resident to resident text occurred at a frequency of 3-5 times/d (43%) compared with most attending to resident texts, which occurred 1-2 times/d (33%). Most resident to attending (25%) and attending to attending (30%) texts occurred 1-2 times/d. Among those that texted, PRC was the most frequently reported purpose for resident to resident (46%), resident to attending (64%), attending to resident (82%), and attending to other attending staff (60%) texting. Texting was the most preferred method to communicate about routine PRC (47% of residents vs 44% of attendings). Age (OR: 0.86, 95% CI: 0.79-0.95; p = 0.003), but not sex, specialty/clinical rotation, academic rank, or postgraduate year (PGY) level predicted PRC texting. CONCLUSIONS: Most resident and attending staff surveyed utilize texting, mostly for PRC. Texting was preferred for communicating routine PRC information. Our data may facilitate the development of guidelines for the appropriate use of PRC texting.

Comparative morbidity and mortality from cervical or thoracic esophageal anastomoses.

BACKGROUND: In the modern era of esophagectomy, we hypothesized that perioperative morbidity and mortality from cervical or thoracic sites of anastomoses would not be different. METHODS: We used the American College of Surgeons National Surgical Quality Improvement Program database to identify patients who underwent esophagectomy for lower esophageal or gastroesophageal (GE) junction malignancies from 2005 to 2010. Patients were categorized as having either a cervical or thoracic anastomosis based on CPT codes. RESULTS: There were 601 (66%) cervical and 308 (34%) thoracic anastomoses. Cervical anastomoses were associated with greater than 2 units of blood transfusion in a higher proportion of patients (10% vs. 3%, P = 0.001), and higher superficial surgical site infections (13% vs. 7%, P = 0.003). There were no difference in rates of organ/space infections (6% vs. 7%, P = 0.70), overall morbidity (38% vs. 39%, P = 0.84), or mortality (3% vs. 4%, P = 0.34). Median length of stay was similar (11.5 days cervical vs. 11 days thoracic, P = 0.89), even among patients with organ/space infections (18 days cervical vs. 21 days thoracic, P = 0.49). On multivariate analysis thoracic anastomosis was not a significant predictor of increased overall morbidity (OR 1.13: 95%CI 0.83-1.54). CONCLUSION: After esophagectomy, the site of anastomosis does not predict an increased risk of perioperative morbidity or mortality.

Age-related disparities in use of completion lymphadenectomy for melanoma sentinel lymph node metastasis.

BACKGROUND: Guidelines recommend that patients with melanoma metastatic to the sentinel lymph node (SLN) undergo a completion lymphadenectomy (CLND) of the affected lymph node basin. We have previously reported on decreased use of SLN biopsy among elderly patients. We hypothesized that elderly patients with SLN metastases would have lower rates of CLND relative to their younger counterparts. METHODS: The Surveillance, Epidemiology, and End Results database was queried for patients who underwent SLN biopsy for intermediate thickness cutaneous melanoma (Breslow thickness 1.01 mm-4.00 mm) from 2004 to 2008 and were found to have SLN metastasis. Patients were categorized according to age by decade. We then used multivariate logistic regression models to predict receipt of CLND. Additional covariates included sex, race/ethnicity, T stage, tumor histology, tumor location, and ulceration. The likelihood of receiving a CLND was reported as OR with 95% CI; significance was set at P ≤ 0.05. RESULTS: Entry criteria were met by 765 patients. Of these, 548 (71.6%) patients underwent CLND. On multivariate analysis, patients in the age groups 70-79 y old (OR 0.39, CI 0.20-0.78; P = 0.007) and ≥ 80 y old (OR 0.27, CI 0.12-0.61; P = 0.001) were less likely to undergo CLND than the youngest age group (1-39 y old). CONCLUSIONS: Elderly patients with SLN metastasis are less likely to receive CLND than their younger counterparts. A multi-center randomized clinical trial evaluating the potential survival benefit of CLND is ongoing. Further research to assess reasons why the elderly are less likely to receive CLND are needed.

Assessing rural-urban disparities in the use of sentinel lymph node biopsy for melanoma.

BACKGROUND: We hypothesized that patients in urban areas with intermediate thickness cutaneous melanoma would have higher rates of sentinel lymph node biopsy (SLNB) relative to their rural-dwelling counterparts. METHODS: The Surveillance, Epidemiology, and End Results database was queried for patients who underwent surgery for intermediate thickness cutaneous melanoma from 2004-2008. Patients were categorized as coming from urban or rural counties based on a nine-point scale. We used multivariate logistic regression models to predict use of SLNB. Covariates examined included sex, race/ethnicity, age, T stage, tumor histology, tumor location, and ulceration. The likelihood of undergoing SLNB was reported as OR with 95% CI. RESULTS: Of 8441 patients, 8382 (99.3%) had complete information regarding use of SLNB. On multivariate analysis, patients from rural counties had a decreased likelihood of receiving a SLNB (OR 0.87, CI 0.78-0.97; P = 0.014). Additional factors associated with a decreased likelihood of receiving a SLNB included increasing age, Asian/Hispanic/Unknown race, and head and neck or overlapping primary tumor site. CONCLUSIONS: Patients in rural areas are less likely to receive a SLNB for intermediate thickness cutaneous melanoma than their urban-dwelling counterparts.

Current oncologic applications of radiofrequency ablation therapies.

Radiofrequency ablation (RFA) uses high frequency alternating current to heat a volume of tissue around a needle electrode to induce focal coagulative necrosis with minimal injury to surrounding tissues. RFA can be performed via an open, laparoscopic, or image guided percutaneous approach and be performed under general or local anesthesia. Advances in delivery mechanisms, electrode designs, and higher power generators have increased the maximum volume that can be ablated, while maximizing oncological outcomes. In general, RFA is used to control local tumor growth, prevent recurrence, palliate symptoms, and improve survival in a subset of patients that are not candidates for surgical resection. It's equivalence to surgical resection has yet to be proven in large randomized control trials. Currently, the use of RFA has been well described as a primary or adjuvant treatment modality of limited but unresectable hepatocellular carcinoma, liver metastasis, especially colorectal cancer metastases, primary lung tumors, renal cell carcinoma, boney metastasis and osteoid osteomas. The role of RFA in the primary treatment of early stage breast cancer is still evolving. This review will discuss the general features of RFA and outline its role in commonly encountered solid tumors.

Sentinel lymph node biopsy in patients with thick primary cutaneous melanoma: patterns of use and underuse utilizing a population-based model.

Background. Sentinel lymph node biopsy (SLNB) for thick cutaneous melanoma is supported by national guidelines. We report on factors associated with the use and underuse of SLNB for thick primary cutaneous melanoma. Methods. The Surveillance, Epidemiology, and End Results database was queried for patients who underwent surgery for thick primary cutaneous melanoma from 2004 to 2008. We used multivariate logistic regression models to predict use of SLNB. Results. Among 1,981 patients, 833 (41.8%) did not undergo SLNB. Patients with primary melanomas of the arm (OR 2.07, CI 1.56-2.75; P < 0.001), leg (OR 2.40, CI 1.70-3.40; P < 0.001), and trunk (OR 1.82, CI 1.38-2.40; P < 0.001) had an increased likelihood of receiving a SLNB, as did those with desmoplastic histology (OR 1.47, CI 1.11-1.96; P = 0.008). A decreased likelihood of receiving SLNB was noted for advancing age ≥ 60 years (age 60 to 69: OR 0.58, CI 0.33-0.99, P = 0.047; age 70 to 79: OR 0.32, CI 0.19-0.54, P < 0.001; age 80 or more: OR 0.10, CI 0.06-0.16, P < 0.001) and unknown race/ethnicity (OR 0.21, CI 0.07-0.62; P = 0.005). Conclusions. In particular, elderly patients are less likely to receive SLNB. Further research is needed to assess whether use of SLNB in this population is detrimental or beneficial.

Negligible effect of perioperative epidural analgesia among patients undergoing elective gastric and pancreatic resections.

BACKGROUND: There are conflicting data regarding improvements in postoperative outcomes with perioperative epidural analgesia. We sought to examine the effect of perioperative epidural analgesia vs. intravenous narcotic analgesia on perioperative outcomes including pain control, morbidity, and mortality in patients undergoing gastric and pancreatic resections. METHODS: We evaluated 169 patients from 2007 to 2011 who underwent open gastric and pancreatic resections for malignancy at a university medical center. Emergency, traumatic, pediatric, enucleations, and disseminated cancer cases were excluded. Clinicopathologic data were reviewed among epidural (E) and non-epidural (NE) patients for their association with perioperative endpoints. RESULTS: One hundred twenty patients (71 %) received an epidural and 49 (29 %) did not. There were no significant differences (P > 0.05) in mean pain scores at each of the four days (days 0-3) among the E (3.2 ± 2.7, 3.2 ± 2.3, 2.3 ± 1.9, and 2.1 ± 1.9, respectively) and NE patients (3.7 ± 2.7, 3.4 ± 1.9, 2.9 ± 2.1, and 2.4 ± 1.9, respectively). Within each of the E and NE patient groups, there were significant differences (P < 0.0001) in mean pain scores from day 0 to day 3 (P < 0.0001). Of the E patients, 69 % also received intravenous patient-controlled analgesia (PCA). Ileus (13 % E vs. 8 % NE), pneumonia (12 % E vs. 8 % NE), venous thromboembolism (6 % E vs. 4 % NE), length of stay [11.0 ± 12.1 (8, 4-107) E vs. 12.2 ± 10.7 (7, 3-54) NE], overall morbidity (36 % E vs. 39 % NE), and mortality (4 % E vs. 2 % NE) were not significantly different. CONCLUSIONS: Routine use of epidurals in this group of patients does not appear to be superior to PCA.

Nomograms to predict risk of in-hospital and post-discharge venous thromboembolism after abdominal and thoracic surgery: an American College of Surgeons National Surgical Quality Improvement Program analysis.

BACKGROUND: Postoperative venous thromboembolism (VTE) is increasingly viewed as a quality of care metric, although risk-adjusted incident rates of postoperative VTE and VTE after hospital discharge (VTEDC) are not available. We sought to characterize the predictors of VTE and VTEDC to develop nomograms to estimate individual risk of VTE and VTEDC. METHODS: Using the American College of Surgeons National Surgical Quality Improvement Program database, we identified 471,867 patients who underwent inpatient abdominal or thoracic operations between 2005 and 2010. We excluded primary vascular and spine operations. We built logistic regression models using stepwise model selection and constructed nomograms for VTE and VTEDC with statistically significant covariates. RESULTS: The overall, unadjusted, 30-d incidence of VTE and VTEDC was 1.5% and 0.5%, respectively. Annual incidence rates remained unchanged over the study period. On multivariate analysis, age, body mass index, presence of preoperative infection, operation for cancer, procedure type (spleen highest), multivisceral resection, and non-bariatric laparoscopic surgery were significant predictors for VTE and VTEDC. Other significant predictors for VTE, but not VTEDC, included a history of chronic obstructive pulmonary disease, disseminated cancer, and emergent operation. We constructed and validated nomograms by bootstrapping. The concordance indices for VTE and VTEDC were 0.77 and 0.67, respectively. CONCLUSIONS: Substantial variation exists in the incidence of VTE and VTEDC, depending on patient and procedural factors. We constructed nomograms to predict individual risk of 30-d VTE and VTEDC. These may allow more targeted quality improvement interventions to reduce VTE and VTEDC in high-risk general and thoracic surgery patients.

Utilization of sentinel lymph node biopsy in patients with ductal carcinoma in situ undergoing mastectomy.

BACKGROUND: Current guidelines suggest consideration of sentinel lymph node biopsy (SLNB) for patients with ductal carcinoma in situ (DCIS) undergoing mastectomy. Our objective was to identify factors influencing the utilization of SLNB in this population. METHODS: We used the Surveillance Epidemiology and End Results database to identify all women with breast DCIS treated with mastectomy from 2000 to 2008. We excluded patients without histologic confirmation, those diagnosed at autopsy, those who had axillary lymph node dissections performed without a preceding SLNB, and those for whom the status of SLNB was unknown. We used multivariate logistic regression reporting odds ratios (OR) and 95% confidence intervals (CI) to evaluate the relationship of patient- and tumor-related factors to the likelihood of undergoing SLNB. RESULTS: Of 20,177 patients, 51% did not receive SLNB. Factors associated with a decreased likelihood of receiving a SLNB included advancing age (OR 0.66; 95% CI 0.62-0.71), Asian (OR 0.75; CI 0.68-0.83) or Hispanic (OR 0.84; 95% CI 0.74-0.96) race/ethnicity, and history of prior non-breast (OR 0.57; 95% CI 0.53-0.61). Factors associated with an increased likelihood of receiving a SLNB included treatment in the east (OR 1.28; 95% CI 1.17-1.4), intermediate (OR 1.25; 95% CI 1.11-1.41), high (OR 1.84; 95% CI 1.62-2.08) grade tumors, treatment after the year 2000, and DCIS size 2-5 cm (OR 1.54; 95% CI 1.42-1.68) and >5 cm (OR 2.43; 95% CI 2.16-2.75). CONCLUSIONS: SLNB is increasingly utilized in patients undergoing mastectomy for DCIS, but disparities in usage remain. Efforts at improving rates of SLNB in this population are warranted.

Treatment options for metaplastic breast cancer.

Metaplastic breast cancer (MBC) is a malignancy characterized by the histologic presence of two or more cellular types, commonly a mixture of epithelial and mesenchymal components. MBC is rare relative to invasive ductal carcinoma (IDC), representing less than 1% of all breast cancers. Other than a lower rate of lymph node metastases, MBC tumors display poorer prognostic features relative to IDC. Due to its low incidence and pathological variability, the ideal treatment paradigm for MBC is unknown. Because of its rarity, MBC has been treated as a variant of IDC. Despite similar treatment regimens, however, patients with MBC have worse outcomes. Recent research is focused on biological differences between MBC and IDC and potential novel targets for chemotherapeutic agents. This paper serves as a summation of current literature on approaches to the multidisciplinary treatment of patients with MBC.

Geographic variation in utilization of sentinel lymph node biopsy for intermediate thickness cutaneous melanoma.

BACKGROUND AND OBJECTIVES: Sentinel lymph node biopsy (SLNB) is the standard for evaluation of the draining lymphatic basin for intermediate thickness melanoma. Despite this, SLNB has not been uniformly adopted. We hypothesized that there are geographic areas of the United States where patients are less likely to receive SLNB. METHODS: The Surveillance, Epidemiology, and End Results database was queried for patients who underwent surgery for intermediate thickness cutaneous melanoma (Breslow thickness 1.00-4.00 mm) from 2004 to 2008. Patients were categorized according to geographic area based on the reporting registry. Multivariate logistic regression models predicted use of SLNB. RESULTS: Entry criteria were met by 8957 patients. On multivariate analysis, patients from the South were less likely (OR 0.54, CI 0.48-0.62; P < 0.001) to receive a SLNB. Additional factors associated with a decreased likelihood of receiving a SLNB included head and neck primary tumor site, high or unknown serum LDH, Asian, Hispanic, Native American or unknown race, and increasing age. CONCLUSIONS: Patients from the South were less likely to receive a SLNB for an intermediate thickness cutaneous melanoma. This report of geographic disparities on a national level should be confirmed locally to better guide interventions aimed at eliminating these disparities.

Rural-urban disparities in use of post-lumpectomy radiation.

We hypothesized that breast cancer (BCa) patients in urban counties would have higher rates of post-lumpectomy radiation therapy (RT) relative to patients in near-metro and rural counties. We used the Surveillance, Epidemiology, and End Results (SEER) database to identify women diagnosed with BCa treated with lumpectomy in the Sacramento area between 2000 and 2006. Patient counties were categorized as urban and near-metro. Multivariate logistic regression models predicted treatment with RT. Likelihood of undergoing RT was reported as odds ratios (OR) with 95 % confidence intervals (CI). Of 7,953 patients meeting entry criteria, 5,858 (73.7 %) underwent RT. On multivariate analysis, patients from near-metro (OR, 0.66; CI, 0.59-0.75; P < 0.001) and rural (OR 0.39, CI 0.30-0.52; P < 0.001) areas had a decreased likelihood of undergoing RT relative to patients from urban areas. Patients from near-metro and rural areas are less likely to receive RT following lumpectomy for BCa than their urban counterparts.

Utilization of lymph node assessment in patients with ductal carcinoma in situ treated with lumpectomy.

BACKGROUND: Lymph node assessment (LNA), including sentinel lymph node biopsy (SLNB), is controversial in patients undergoing lumpectomy for ductal carcinoma in situ (DCIS). Our goal was to identify factors influencing LNA in these patients. METHODS: We used the Surveillance Epidemiology and End Results database to identify all female patients treated with lumpectomy for DCIS from 2000 to 2008. We excluded patients without histologic confirmation, including those diagnosed at autopsy, and those for whom LNA status was unknown. Multivariate logistic regression models predicted use of LNA. Likelihood of undergoing LNA was reported as odds ratios (ORs) with 95% confidence intervals (CIs). RESULTS: A total of 62,935 patients met inclusion criteria. Approximately 15% (N = 9726) had regional LNA at the time of lumpectomy, with 12% (N = 7294) undergoing SLNB. Factors associated with an increased likelihood of undergoing LNA included treatment in the Southeast (OR 1.25, CI 1.04-1.22); treatment after the year 2000; grade II (OR 2.71, CI 2.48-2.96), III (OR 2.38, CI 2.18-2.59), or IV (OR 2.61, CI 2.37-2.88) tumors; DCIS size 2-5 cm (OR 1.49, CI 1.37-1.62) or >5 cm (OR 2.16, CI 1.78-2.61), and estrogen receptor-negative (OR 1.29, CI 1.16-1.43) or progesterone receptor-negative (OR 1.22, CI 1.11-1.33) tumors. Factors associated with a decreased likelihood of undergoing regional LNA were age >60 (OR 0.83, CI 0.79-0.87), and Asian race (OR 0.88, CI 0.81-0.96). Factors predictive of LNA in general were also predictive of SLNB. CONCLUSIONS: Although LNA is controversial for patients undergoing lumpectomy for DCIS, it is used in 15% of cases. Further research establishing for the benefit of LNA in DCIS patients treated with lumpectomy is needed.

Urban and non-urban disparities in the use of post-mastectomy radiation for breast cancer.

Post-mastectomy radiation therapy (PMRT) is indicated for local-regionally advanced breast cancer (LABC). We hypothesized that candidates for PMRT from non-urban areas would receive lower rates of RT than urban patients and would have poorer overall survival (OS) and disease-specific survival (DSS). We used the Surveillance, Epidemiology, and End Results database to identify patients diagnosed with LABC and treated with mastectomy in Sacramento and its surrounding 13 counties between 2000 and 2006. All patients were eligible to receive RT according to established guidelines, with tumors >5 cm size, ≥ 4 metastatic lymph nodes, or both. According to a United States Department of Agriculture scale, we designated counties as urban or non-urban and used multivariate logistic regression and Cox proportional hazards models to predict the use of RT, overall survival (OS), and disease-specific survival (DSS). Density of radiation oncologists in non-urban and urban counties was determined using the American Medical Association database in relation to census-derived populations of the respective counties. Entry criteria were met by 1,507 patients. Most (56.5%) were from urban counties; only 61% received RT. There was no radiation oncologist listed for 8/10 non-urban counties and 2/4 urban counties. Each radiation oncologist served 88,804 people in non-urban counties and 68,624 residents in urban counties. On multivariate analysis, non-urban patients (OR 0.56, CI 0.44-0.72) and increasing age were the only factors predicting a decreased likelihood of receiving RT (OR 0.97, CI 0.96-0.98). Patients not receiving PMRT experienced poorer OS (HR 1.77, CI 1.39-2.25; P < 0.001) and DSS (HR 1.62, CI 1.23-2.15; P = 0.001); however, non-urban status did not predict OS or DSS. Non-urban residents with LABC are less likely to receive indicated PMRT. This discrepancy may be due to limited RT access in non-urban areas. The lack of poorer OS and DSS due to this disparity requires further study.