The effect of a paradigm shift towards minimally invasive surgery on hospitalization costs of hysterectomy for endometrial cancer

Abstracts / Gynecologic Oncology 125 (2012) S3–S167

Michigan State University, Grand Rapids, MI, 2The Ohio State University, Columbus, OH, 3University of Kentucky Medical Center, Lexington, KY, 4 University of Kentucky Medical Center Biostatistical Core, Lexington, KY. Objective: Recent trials concluded that pelvic (PV) lymphadenectomy (LND) for endometrial cancer does not change overall survival. Our study aim was to evaluate survival and adjuvant therapy for women undergoing selective PV&aortic (A) sampling versus systematic PV&ALND. Methods: A retrospective analysis was performed using 2 institutional database pts with epithelial endometrial carcinoma undergoing surgery from 1997 to 2008. Pts were dichotomized according to type of LND: 1 — selective PV and A sampling or PV sampling only, and 2 — systematic PV and ALND or PVLND and A sampling. Demographic, clinicopathologic, Charlson comorbidity index, treatment variables and survival were recorded. Pts were categorized as low, intermediate (I), or high-risk (H) of recurrence. Correlation was assessed using Chi-square. The KM method and Cox PH regression were employed for survival analysis (progression or death, death any cause, death from disease). P-values b0.05 were considered significant. Results: Five hundred and sixty-seven pts met inclusion criteria (Group 1: 252, 2: 315). The mean age & % LNs+ were similar between the two groups. The BMI, Charlson co-morbidity index, stage, histology, median #LNs removed, and adjuvant therapy were different and adjusted for in the analysis. Low-risk pts had a high 5-year DSS regardless of LND procedure. On multivariable analysis adjusted for age, obesity, comorbidity index, histology, LN metastasis, and adjuvant therapy, the type of LND was not associated with survival (aHR: 0.55, 95%CI: 0.27,1.14); however fewer I or H pts relapsed or died in Group 2 compared to 1 (Table 1; HR progression or death: 0.49, 95%CI: 0.26,0.91, P = 0.02). Exploratory analyses comparing a systematic PV&ALND to any other type of dissection resulted in similar findings. Among the intermediate risk pts receiving less than a systematic LND, the odds of receiving adjuvant therapy decreased by 7% for each additional LN removed (aOR 0.93, P = 0.03; adjusted for histology and grade). For each unit increase in the Charlson comorbidity index, the risk of progression and/or death increased by 20% (both HR: 1.2 P = 0.02 and 0.03, respectively). Conclusions: Systematic LND did not affect patient survival, but type of LND did influence the likelihood of adjuvant therapy among the intermediate-risk group receiving less than a systematic PV&ALND. The Charlson co-morbidity index increases the risk of progression and/or death and should be considered for future trials evaluating PFS and OS.

Table 1 Overall survival, disease specific, and recurrence-free survival according to type of lymphadenectomy and risk of recurrence. Low risk A Overall survival Died 6 (3.6%) 3 years 91.8% 5 years 88.9% Disease-specific survival Died 1 (0.6%) 3 years 100.0% 5 years 96.9% Progression-free survival Relapsed or died 7 (4.2%) 3 years 89.5% 5 years 89.5%

P-value A

B

27 (18.2%) 9 (9.2%) 82.5% 95.4% 72.9% 86.5%

0.11

2 (1.3%) 100.0% 100.0%

21 (14.2%) 5 (5.1%) 85.2% 96.6% 80.1% 93.8%

0.05

7 (4.6%) 0.56 92.8% 92.8%

doi:10.1016/j.ygyno.2011.12.045

Objective: To determine the sensitivity and specificity of fluorescence imaging of indocyanine green (ICG) and standard colorimetric analysis of isosulfan blue (ISB) dye for detecting sentinel lymph nodes (SLN) in women with endometrial cancer (EC) undergoing robotic-assisted lymphadenectomy (RAL). Secondary aim was to investigate the ability of SLN analysis to increase the detection of metastatic disease by comparing SLN immunohistochemical (IHC) and hematoxolin and eosin (H&E) results. Methods: Thirty-five patients underwent robotic hysterectomy with RAL for EC. 1 mL ISB was injected sub-dermally in four quadrants of the cervix, followed by 0.5 mL of diluted ICG (25 mg in 20 mL saline) immediately prior to placement of a uterine manipulator. Paravesical and pararectal spaces were dissected for colorimetric detection of lymphatic pathways and SLN. The da Vinci® camera was switched to Firefly mode and fluorescent imaging results were recorded. SLN were removed for permanent analysis with ultra-sectioning, H&E, and IHC staining. Results: Twenty-seven (77%) and 34 (97%) of patients had bilateral pelvic or aortic SLN detected by colorimetric and fluorescence, respectively (p = 0.03). Considering each hemi-pelvis separately, 15/70 (21.4%) had “weak” uptake of ISB in SLN confirmed positive with fluorescence imaging. All patients had at least unilateral SLN by each method. 10 (28.6%) patients had lymph node (LN) metastasis, and 9 of these had SLN metastasis (90% sensitivity, one false negative SLN biopsy). 25 patients had normal LN, all with negative SLN biopsies (100% specificity). Four (40%) of patients with LN metastasis were detected only by IHC and ultra-sectioning of SLN (67% increase in detection of metastasis using SLN). Twentytwo (63%) patients had comprehensive pelvic and infra-renal lymphadenectomy and 13 had pelvic dissections only. For all cases, mean operative time was 169 ± 37 min, EBL 118 ± 47 mL, BMI 33.1 ± 9.3 kg/m2, lesion size 4.4 ± 2.3 cm, invasion 41 ± 34%, LVSI 40%, G2/G3 63%, pelvic LN 22.6 ± 10.9, and aortic LN 10.3 ± 6.6. Conclusions: Fluorescence imaging with ICG detected bilateral SLN more often than ISB colorimetric analysis, and fluorescence imaging also confirmed SLN in patients with weak nodal uptake of ISB. SLN mapping with IHC staining increased the detection of lymph node metastasis. A larger cohort will be necessary to confirm sensitivity and specificity of fluorescence imaging of SLN, but this study supports the growing body of literature advocating the use of SLN mapping for patients with EC.

P-value

5 (3.3%) 0.27 97.8% 97.8% 0.88

45 Detection of sentinel lymph nodes in patients with endometrial cancer undergoing robotic-assisted staging: A comparison of colorimetric and fluorescence imaging R. Holloway, R. Molero Bravo, J. Rakowski, J. James, C. Jeppson, S. Ingersoll, S. Ahmad. Florida Hospital Cancer Institute, Orlando, FL.

doi:10.1016/j.ygyno.2011.12.046

Intermediate or high risk B

S19

40 (27.2%) 13 (13.5%) 0.02 69.1% 86.2% 53.2% 71.5%

46 The effect of a paradigm shift towards minimally invasive surgery on hospitalization costs of hysterectomy for endometrial cancer N. Shah1, C. Akileswaran2, L. Garrett3, L. Bradford3, M. Del Carmen3, A. Goodman3, J. Schorge3, D. Boruta3, W. Growdon3. 1Massachusetts General Hospital, Boston, MA, 2University of California, San Francisco, San Francisco, CA, 3Massachusetts General Hospital/Harvard University, Boston, MA. Objective: Our hospital experienced a seven-fold increase in the rate of minimally invasive surgery (MIS) for endometrial cancer between 2006 and 2010. We sought to determine the effect of this shift on the cost of hospitalization for upfront endometrial cancer surgery.

S20

Abstracts / Gynecologic Oncology 125 (2012) S3–S167

Methods: We performed a retrospective cohort analysis of patient characteristics, operative times, complication rates, and costs from 452 consecutive cases of patients who underwent hysterectomy for endometrial cancer by any method between January 2006 and March 2010 at our hospital. All costs, charges, and payments of hospitalization associated with robotic, abdominal, and laparoscopic methods of hysterectomy were examined from the hospital perspective, and obtained directly from the hospital accounting ledgers. All monetary values were converted to real 2011 dollars by applying inflation rates based on the United States Consumer Price Index. We compared differences in cost assessment, charges, and reimbursement for the period of hospitalization across this time interval using χ2, Wilcoxon ranksum, student t tests. Factors associated with increased cost were assessed using both linear and logistic regression. Results: The rate of MIS increased from 8.9% in 2006 to 63% in 2010 (p b 0.001). No differences in age, stage distribution, major surgical complications, or conversion rate to laparotomy were observed across this time period. An increased rate of comprehensive surgical staging (66% to 84%, p b 0.002) and total operating room utilization time (189 to 281 min, p b 0.001) was observed from 2006 to 2010. Over this period, median length of stay decreased from 3 to 2 days (p b 0.003). There was no significant difference in the average inflation-adjusted cost assessed by the hospital over the time interval (p = 0.355). While the total charges per patient increased (p b 0.02), no difference was observed in average reimbursed payments over time (p = 0.447). Multivariate analysis identified increased operating room utilization time, longer length of stay, and decision to perform MIS as independent factors associated with increased cost assessment. Conclusions: In the near term a rapid shift toward MIS at our institution did not cause a significant change in the average cost of hospitalization. This may be because the cost savings from decreased length of stay are offset by increase in operating room utilization time as surgeons newly adopted MIS technology and took on increasingly complex cases.

Year Average cost Mean operating room time (min) Median length of stay (days)

2006 $13,211 188 3

2007 $16,047 197 3

2008 $14,710 228 2

2009 $16,017 287 2

47 Incidence of venous thromboembolism after minimally invasive surgery in patients with newly diagnosed endometrial cancer S. Sandadi, A. Walter, S. Lee, G. Gardner, N. Abu-Rustum, Y. Sonoda, C. Brown, E. Jewell, M. Leitao. Memorial Sloan-Kettering Cancer Center, New York, NY. Objective: To report the incidence of postoperative venous thromboembolism (VTE) among patients undergoing minimally invasive surgery (MIS) for endometrial cancer, and to characterize risk factors associated with the development of VTE. Methods: All patients with newly diagnosed endometrial cancer who were scheduled to undergo a planned MIS procedure from 5/1/07 to 12/31/10 were identified and data were abstracted from the electronic medical records. The incidence of symptomatic postoperative VTE was determined in the patients that did not require conversion to laparotomy. Various clinicopathologic variables were tested for an association with the development of postoperative VTE using standard statistical tests. Results: We identified 573 cases that did not require conversion to laparotomy: 310 (54%) were completed robotically (RBT) and 263 (46%) were completed with standard laparoscopy (LSC). Postoperative low molecular weight heparin (LMWH) was administered to 125 (22%) patients during their stay. All patients had sequential compression devices (SCD) placed intraoperatively. Seven (1.2%) patients developed a symptomatic VTE: 2 with a DVT only, and 5 with a PE without an identified DVT. The factors associated with development of a postoperative VTE were: BMI (P = 0.005), estimated blood loss (P = 0.03), and operative time (P = 0.01). Median length of postoperative stay was also associated with VTE development (P = 0.005). The use of postoperative anticoagulation, stage, tumor grade, and tumor histology was not associated with VTE development. In patients with a BMI ≥ 40 kg/m2, the incidence of VTE was 5.8% (4/69), compared to 0.6% (3/504) in patients with BMI b 40 kg/m2 (P = 0.005). A very high risk group was determined to be patients with a BMI ≥ 40 kg/m2 and an operative time ≥180 min. In this very high risk group, the incidence of VTE was 9.5% (4/42), compared to 0.6% (3/531) in all others (P = 0.001). Postoperative LMWH was used in 19/42 (45%) very high risk cases. There were no postoperative deaths. Conclusions: The incidence of VTE in patients with newly diagnosed endometrial cancer undergoing successful MIS approaches is very low. There appears to be no clear justification for the routine use of a heparin for perioperative thromboprophylaxis in the majority of these patients. Thromboprophylaxis with heparin may, however, be a consideration in morbidly obese patients (BMI ≥ 40), especially after a procedure that lasts ≥3 h. Further investigation is warranted.

doi:10.1016/j.ygyno.2011.12.048

48 Reduction in venous thromboembolism (VTE) rates following the implementation of extended duration prophylaxis for patients undergoing surgery for gynecologic malignancies K. Schmeler, G. Langley, K. Cain, M. Munsell, P. Ramirez, P. Soliman, A. Nick, M. Frumovitz, E. Garcia, C. Levenback. The University of Texas, MD Anderson Cancer Center, Houston, TX. doi:10.1016/j.ygyno.2011.12.047

Objective: Current guidelines recommend extended duration pharmacologic prophylaxis for the prevention of venous thromboembolism (VTE) in women undergoing pelvic surgery for cancer. The objective of this study was to compare the incidence of VTE