Abstract
Background and Objectives:
Recent studies demonstrating decreased survival following minimally invasive surgery (MIS) for cervical cancer have generated concern regarding oncologic efficacy of MIS. Our objective was to evaluate the association between surgical approach and 5-year survival following resection of abdominopelvic malignancies.
Methods:
Patients with stage I or II adenocarcinoma of the prostate, colon, rectum, and stage IA2 or IB1 cervical cancer from 2010-2015 were identified from the National Cancer Data Base. The association between surgical approach and 5-year survival was assessed using propensity-score-matched cohorts. Distributions were compared using logistic regression. Hazard ratio for death was estimated using Cox proportional-hazard models.
Results:
The rate of deaths at 5 years was 3.4% following radical prostatectomy, 22.9% following colectomy, 18.6% following proctectomy, and 6.8% following radical hysterectomy. Open surgery was associated with worse survival following radical prostatectomy (HR, 1.18; 95% CI, 1.05-1.33; P = .005), colectomy (HR, 1.45; 95% CI, 1.39-1.51; P < .001), and proctectomy (HR, 1.28; 95% CI, 1.10-1.50; P = .002); however, open surgery was associated with improved survival following radical hysterectomy (HR, 0.61; 95% CI, 0.44-0.82; P = .003).
Conclusions:
These results suggest that MIS is an acceptable approach in selected patients with prostate, colon, and rectal cancers, while concerns regarding MIS resection of cervical cancer appear warranted.
Keywords: abdominopelvic malignancies, minimally invasive surgery, survival
1 |. INTRODUCTION
Following the first reported laparoscopic surgical procedures in the late 1980s and early 1990s, there has been a rapid adoption of minimally invasive surgery (MIS) throughout many surgical specialties.1,2 This diffusion has substantially impacted the approach to surgical management of abdominopelvic malignancies, with minimally invasive techniques (including both traditional laparoscopic and robotically-assisted approaches) comprising a large and increasing share of the surgical volume for many cancer types.3–7
Despite the expansion of MIS techniques, concerns persist regarding the level of evidence justifying their use for oncologic indications. There are relatively few prospective randomized comparisons of surgical approach and most of those studies were carried out at pioneering centers early in the MIS era.8–11 Furthermore, many studies have compared peri-operative safety outcomes with fewer studies examining longer-term oncologic outcomes.12–16 Less is known about outcomes in broader practice when this technology extends beyond the confines of selected high-volume centers, and some have raised concerns that the adoption of MIS techniques may have, at times, outpaced the rigorous research needed to completely evaluate effectiveness.17–20
Recently, these concerns have been augmented in the context of two studies demonstrating worse overall survival amongst early stage cervical cancer patients who underwent MIS compared to open surgery.21,22 There has been significant media publicity following these reports, leading the US Food and Drug Administration to issue a warning against the use of minimally invasive approaches for certain malignancies pending further evaluation.23 The objectives of this study were (1) to evaluate the association between surgical approach and 5-year survival in patients with early stage prostate, colon, rectal, and cervical cancers and (2) to assess other operative outcomes associated with surgical approach.
2 |. METHODS
2.1 |. Data source
The National Cancer Data Base (NCDB) is a registry sponsored by the American College of Surgeons, the Commission on Cancer, and the American Cancer Society. Information regarding data abstraction, sampling strategy, variables collected and outcomes are detailed within the literature.24,25 Briefly, the NCDB began collecting information on incident malignancies in 1989, includes approximately 70% of newly diagnosed cancers from over 1500 hospitals within the United States, and currently contains more than 34 million historical records.26 Additional information regarding the NCDB is provided in the eMethods. This study was deemed exempt by the Institutional Review Board of Northwestern University.
2.2 |. Study population
Data from the NCDB were used to identify patients diagnosed with either stage I or II prostate, and colon adenocarcinoma, as well as stage I rectal adenocarcinoma from 2010 to 2015 and stage IA2 or IB1 cervical cancer from 2010 to 2015 for this retrospective cohort study. Early stage cervical cancer was included in the analyses to confirm reproducibility of findings in recent studies, as well as to affirm the methodology for prostate, colon, and rectal cancers. Early stage prostate, colon, and rectal cancers were included to evaluate operative resection of common pelvic and non-pelvic malignancies. Patients were identified using the International Classification of Diseases for Oncology, third edition, site and histology codes.27 Adenocarcinoma codes (8140, 8141, 8142, 8143, 8144, 8145, 8147, 8480, 8481, and 8190) were used to identify prostate, colon, and rectal cancer patients. Adenocarcinoma, adenosquamous carcinoma (8560 and 8570) and squamous-cell carcinoma codes (8070, 8071, 8072, 8073, 8074, 8075, 8076, and 8078) were used to identify cervical cancer patients. Surgical resection was defined as radical prostatectomy, colectomy, proctectomy, or radical hysterectomy. Patients were excluded if surgical approach was unknown, they had more than one or unknown number of primary tumors, had no pathologic confirmation of disease. The surgical approach variable was used to determine patients who underwent open, laparoscopic, or robot-assisted surgery. In this analysis, patients were classified as having MIS if they underwent laparoscopic or robot-assisted surgery. The remaining patients, including those who were converted to open surgery, were categorized as open surgery. Further information regarding definitions for inclusion and exclusion criteria is provided in the eMethods.
2.3 |. Outcomes
The primary outcome of interest was overall 5-year survival. Secondary outcomes of interest included readmission within 30 days postoperatively, surgical margin status, 30-day mortality, and 90-day mortality.
2.4 |. Covariates
Covariates included demographic, socioeconomic, and clinical variables within NCDB. These included sex, age, race (White, Black, Hispanic, Other/Unknown), insurance type, income, education, rural or urban status, Charlson Index, hospital type, serum prostate-specific antigen (PSA), Gleason score, and stage of disease. Insurance status was categorized as uninsured, private, Medicaid, and Medicare. The proportion of patients without a high-school diploma along with median income were obtained using ZIP Code-level estimates. The U.S Department of Agriculture 2013 Rural-Urban Continuum Codes classification was used to categorize rural or urban status as metropolitan, metropolitan adjacent, and rural.28 Patients with one or more coexisting medical conditions were identified using the Charlson Comorbidity Index as recorded in the NCDB.29 Hospital type was classified using the Commission on Cancer criteria as academic or community. Clinical stage of disease for prostate, colon, and rectal cancer was determined using the American Joint Committee on Cancer (AJCC), seventh edition.30 AJCC pathologic stage of disease was used when AJCC clinical stage was unknown. Cervical cancer stage was determined using AJCC clinical stage, Collaborative Stage Site-Specific Factor 1 (Federation of Gynecology and Obstetrics system for cervical cancer) when clinical stage was unknown, and AJCC pathologic stage when both clinical stage and Site-Specific Factor 1 were unknown.31 Clinical stage was used as the primary determinant of disease stage, rather than pathologic stage, as it was available to surgeons in the preoperative phase when decisions regarding surgical approach are made.
2.5 |. Statistical analysis
Propensity score matching was used to construct a cohort of patients who underwent MIS matched with patients who underwent open surgery.32–34 The propensity for each patient to undergo MIS was estimated using a multivariable logistic regression model that regressed receipt of MIS on confounding factors for which we had data. These confounders included sex, age, race, facility type, geographic location, insurance, year of procedure, rural or urban status, income, education, stage of disease, and Charlson Comorbidity Index. Patients were matched in a one-to-one fashion, based on their assigned score, with a caliper diameter of 0.005. Common support following propensity score matching was evaluated graphically to ensure appropriate fit.33
Differences in patient characteristics were evaluated using χ2 tests for categorical variables and two-sided t tests for continuous variables. We examined differences in outcomes between patients receiving MIS vs open surgery. We used the Kaplan-Meier estimator of survival to plot 5-year survival and used the log-rank test to compare differences in survival across MIS and open groups.35 Cox proportional-hazard models were used to estimate the risk of death due to any cause after open surgery (vs MIS) at 5 years. Logistic regression models were estimated to assess the association between surgical approach and readmission 30 days postoperatively, as well as 30- and 90-day mortality by surgical approach. All analyses were performed on the propensity-score-matched sample with robust standard errors to account for clustering of patients within hospitals.
Sensitivity analyses were performed to evaluate the robustness of the results. First, we repeated the analyses using inverse probability of treatment propensity score weighting to ensure that results were not altered by using alternative analytic strategies. In this approach, we used the same propensity scores to weight each observation by (1/(1–propensity score)) if a patient underwent open surgery, and by (1/propensity score) if a patient underwent MIS.36,37 Next, to assess robustness of survival analyses and mitigate the impact of perioperative mortality, conditional survival analyses were performed comparing outcomes amongst patients who survived the peri-operative period (ie, >90 days after surgery).38 In addition, survival analyses were repeated with patients who underwent conversion to open surgery grouped within the MIS cohort rather than the open cohort. Finally, to evaluate adjuvant therapy as a potential confounder of treatment-related survival differences, an adjuvant therapy covariate (no treatment, chemotherapy, chemoradiotherapy, radiotherapy) was added to each survival model. Additional information regarding sensitivity analyses are included in the eMethods. Statistical analysis was performed using STATA v15.1 (College Station, TX).
3 |. RESULTS
3.1 |. Study population
In the matched sample, 76,045 patients underwent radical prostatectomy at 1145 hospitals, 60,553 patients underwent colectomy at 1311 hospitals, 4925 patients underwent proctectomy at 991 hospitals, and 2289 patients underwent radical hysterectomy at 433 hospitals (Figure 1). Of these patients, 40,045 (52.7%) underwent MIS radical prostatectomy, 29,408 (48.6%) underwent MIS colectomy, 2094 (42.5%) underwent MIS proctectomy, and 1109 (48.5%) underwent MIS radical hysterectomy. Patient demographics following propensity-score matching are summarized in Table 1. Each of the covariates were well balanced following the propensity-score match. Patient demographics before and after propensity-score matching are summarized in Tables S1 and S2 in the Appendix.
FIGURE 1.
Study population for prostate, cervical, colon, and rectal cancers [Color figure can be viewed at wileyonlinelibrary.com]
TABLE 1.
Patient characteristics following propensity-score matching
| No. (%) |
||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Prostate |
Colon |
Rectal |
Cervical |
|||||||||
| Open | MIS | P value | Open | MIS | P value | Open | MIS | P value | Open | MIS | P value | |
| Total | 36 000 (47.3) | 40 045 (52.7) | – | 31 145 (51.4) | 29 408 (48.6) | – | 2831 (57.5) | 2094 (42.5) | – | 1180 (51.5) | 1109 (48.5) | – |
| Age, mean, y (SD) | 61.4 (7.0) | 61.5 (7.1) | .570 | 69.4 (12.5) | 69.3 (12.2) | .616 | 64.4 (12.4) | 64.4 (12.1) | .974 | 45.1 (11.7) | 45.1 (11.8) | .976 |
| Race | ||||||||||||
| White | 28 226 (78.4) | 30 031 (75.0) | .651 | 24 782 (79.6) | 23 283 (79.2) | .996 | 2329 (82.2) | 1717 (82.0) | .979 | 750 (63.7) | 733 (66.1) | .959 |
| Black | 4910 (13.6) | 6304 (15.7) | 3324 (10.7) | 3142 (10.7) | 169 (6.0) | 117 (5.6) | 132 (11.2) | 1109 (9.8) | ||||
| Hispanic | 1803 (5.0) | 2357 (5.9) | 1575 (5.1) | 1570 (5.3) | 164 (5.8) | 127 (6.1) | 203 (17.2) | 175 (15.8) | ||||
| Other/unknown | 1061 (2.9) | 1353 (3.4) | 1464 (4.7) | 1413 (4.8) | 169 (6.0) | 133 (6.4) | 95 (8.1) | 92 (8.3) | ||||
| Insurance | ||||||||||||
| Uninsured | 520 (1.4) | 808 (2.0) | .09 | 517 (1.7) | 500 (1.7) | .994 | 47 (1.7) | 42 (2.0) | .869 | 101 (8.6) | 79 (7.1) | .786 |
| Private | 22,309 (62.0) | 23 864 (59.6) | 10,294 (33.1) | 9742 (33.1) | 1343 (47.4) | 985 (47.0) | 687 (58.2) | 657 (59.2) | ||||
| Medicaid | 1035 (2.9) | 1384 (34.9) | 1152 (3.7) | 1102 (3.8) | 141 (5.0) | 102 (4.9) | 245 (20.8) | 237 (21.4) | ||||
| Medicare | 12,136 (33.7) | 13 989 (34.9) | 19,182 (61.6) | 18 064 (61.4) | 1300 (45.9) | 965 (46.1) | 147 (12.5) | 136 (12.3) | ||||
| Income | ||||||||||||
| <38k | 5527 (15.4) | 6934 (17.3) | .597 | 4920 (15.8) | 4676 (15.9) | .975 | 423 (14.9) | 305 (14.6) | .937 | 248 (21.0) | 221 (19.9) | .951 |
| 38-48k | 8042 (22.3) | 9384 (23.4) | 7072 (22.7) | 6788 (23.1) | 675 (23.8) | 507 (24.2) | 297 (25.2) | 273 (24.6) | ||||
| 48-63k | 10,243 (28.5) | 11673 (29.2) | 8596 (27.6) | 8242 (28.0) | 772 (27.3) | 599 (28.6) | 332 (28.1) | 333 (30.0) | ||||
| >63k | 12,188 (33.9) | 12 054 (30.1) | 10 557 (33.9) | 9702 (33.0) | 961 (34.0) | 683 (32.6) | 303 (25.7) | 282 (25.4) | ||||
| Year of diagnosis | ||||||||||||
| 2010 | 9531 (26.5) | 10 774 (26.9) | .998 | 3518 (11.3) | 3138 (10.7) | 0.938 | 187 (6.6) | 150 (7.2) | .969 | 154 (13.1) | 135 (12.2) | .907 |
| 2011 | 8303 (23.1) | 8994 (22.5) | 4737 (15.2) | 4504 (15.3) | 463 (16.4) | 323 (15.4) | 265 (22.5) | 238 (21.5) | ||||
| 2012 | 5661 (15.7) | 6449 (16.1) | 5667 (182) | 5366 (18.3) | 550 (19.4) | 404 (19.3) | 249 (21.1) | 223 (20.1) | ||||
| 2013 | 5124 (14.2) | 5718 (14.3) | 5849 (18.8) | 5618 (19.1) | 578 (20.4) | 419 (20.0) | 222 (18.8) | 216 (19.5) | ||||
| 2014 | 4140 (11.5) | 4547 (11.4) | 5,798 (18.6) | 5524 (18.8) | 543 (19.2) | 404 (19.3) | 200 (17.0) | 197 (17.8) | ||||
| 2105 | 3241 (9.0) | 3563 (8.9) | 5576 (17.9) | 5258 (17.9) | 510 (18.0) | 394 (18.8) | 90 (7.6) | 100 (9.0) | ||||
| Charlson score | ||||||||||||
| 0 | 29 235 (81.2 | 31:644 (19.0) | .090 | 20 334 (65.3) | 19 141 (65.1) | .995 | 2075 (73.3) | 1534 (73.3) | .625 | 1025 (86.9) | 962 (86.7) | .979 |
| 1 | 5768 (16.0) | 7100 (17.7) | 7541 (24.2) | 7164 (24.4) | 575 (20.3) | 405 (19.3) | 124 (10.5) | 118 (10.6) | ||||
| 2 | 813 (2.3) | 1078 (2.7) | 2269 (7.3) | 2157 (7.3) | 126 (4.5) | 104 (5.0) | 22 (1.9) | 22 (2.0) | ||||
| 3 | 184 (0.5) | 223 (0.6) | 1001 (3.2) | 946 (3.2) | 55 (2.9) | 52 (2.4) | 9 (0.8) | 7 (0.6) | ||||
| Histology | ||||||||||||
| Adenocarcinoma | 36 000 (47.3) | 40 045 (52.7) | – | 31145 (51.4) | 29 408 (48.6) | – | 2831 (57.5) | 2094 (42.5) | – | 384 (32.5) | 376 (33.9) | .809 |
| Squamous | – | – | – | – | – | – | 731 (62.0) | 671 (60.5) | ||||
| Adenosquamous | – | – | – | – | – | – | 65 (5.5) | 62 (5.6) | ||||
| Stage | ||||||||||||
| 1 | 12 416 (34.5) | 13 494 (33.7) | .166 | 13 625 (43.8) | 12,731 (43.3) | .934 | 1828 (64.6) | 1377 (65.8) | .762 | – | – | .875 |
| 2A | 15 287 (42.5) | 16 345 (40.8) | 16 179 (52.0) | 15,372 (52.3) | 968 (34.2) | 687 (32.8) | – | – | ||||
| 2B | 8297 (23.1) | 10 206 (25.5) | 1195 (3.8) | 1,175 (4.0) | 33 (1.2) | 27 (1.3) | – | – | ||||
| 2C | – | – | 146 (0.5) | 130 (0.4) | 2 (0.1) | 3 (0.1) | – | – | ||||
| 1A2 | – | – | – | – | – | – | 451 (38.2) | 419 (37.8) | ||||
| 1B1 | – | – | – | – | – | – | 729 (61.8) | 690 (62.2) | ||||
| PSA | ||||||||||||
| 0-10 | 26 139 (72.6) | 27 682 (69.1) | .491 | – | – | – | – | – | – | – | – | – |
| 11-20 | 3886 (10.8) | 4815 (12.0) | – | – | – | – | – | – | ||||
| >20 | 2228 (6.2) | 2837 (7.1) | – | – | – | – | – | – | ||||
| Unknown | 3747 (10.4) | 4711 (11.8) | – | – | – | – | – | – | ||||
| Gleason | ||||||||||||
| <7 | 13 665 (38.0) | 14 946 (37.3) | .412 | – | – | – | – | – | – | – | – | – |
| 7 | 1240 (42.3) | 16 275 (40.6) | – | – | – | – | – | – | ||||
| 8-10 | 4973 (13.8) | 6103 (15.2) | – | – | – | – | – | – | ||||
| Unknown | 2122 (5.9) | 2721 (6.8) | – | – | – | – | – | – | ||||
3.2 |. Survival analysis
The median follow-up in the matched cohorts was 51 months for prostate cancer, 35 months for colon cancer, 36 months for rectal cancer, and 44 months for cervical cancer. The total number of deaths at 5 years were 2550 (3.4%) following radical prostatectomy, 11 884 (22.9%) following colectomy, 790 (18.6%) following proctectomy, and 155 (6.8%) following radical hysterectomy (Table 2). Patients who underwent open surgery had worse 5-year survival when compared to those who underwent MIS for prostate (96.4% vs 96.9%; HR, 1.18; 95% CI, 1.05-1.3; P = .005), colon (73.6% vs 80.8%; HR, 1.45; 95% CI, 1.39-1.51; P < .001) and rectal (79.8% vs 83.7%; HR, 1.28; 95% CI, 1.10-1.50; P = .002) cancer. Conversely, patients who underwent open radical hysterectomy for early stage cervical cancer had significantly improved 5-year survival when compared to those who underwent MIS (94.7% vs 91.4%; HR, 0.61; 95% CI, 0.44-0.82; P = .003). Survival curves stratified by surgical approach are plotted in Figure 2.
TABLE 2.
Unadjusted rates of survival and postoperative outcomes by disease site
| No. (%) |
||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Prostate |
Colon |
Rectal |
Cervical |
|||||||||
| Open | MIS | P value* | Open | MIS | P value* | Open | MIS | P value* | Open | MIS | P value* | |
| 5-y Survival | 34 707 (96.4) | 38 788 (96.9) | .059 | 19 743 (73.6) | 20 261 (80.8) | <.001 | 1971 (79.8) | 1494 (83.7) | .003 | 1119 (94.7) | 1016 (91.4) | .003 |
| 30-d Mortality | 67 (0.2) | 57 (0.1) | .325 | 1147 (3.7) | 506 (1.7) | <.001 | 58 (2.1) | 26 (1.2) | .055 | 1 (0.1) | 2 (0.2) | .528 |
| 90-d Mortality | 92 (0.3) | 89 (0.2) | .572 | 1782 (5.7) | 804 (2.7) | <.001 | 102 (3.6) | 48 (2.3) | .023 | 3 (0.3) | 2 (0.2) | .705 |
| Readmission at 30-d | 1095 (3.0) | 1109 (2.8) | .317 | 2117 (6.8) | 1739 (5.9) | .431 | 233 (8.2) | 180 (8.6) | .136 | 71 (6.0) | 66 (6.0) | .997 |
| Margins | ||||||||||||
| Negative | 27 301 (75.8) | 30 942 (77.3) | .047 | 30 219 (97.0) | 28 891 (98.2) | <.001 | 2708 (95.7) | 2024 (96.7) | .479 | 1114 (94.4) | 1048 (94.5) | .336 |
| Microscopic (R1) | 8360 (23.2) | 8806 (22.0) | 806 (2.6) | 440 (1.5) | 106 (3.7) | 63 (3.0) | 38 (3.2) | 48 (4.3) | ||||
| Macroscopic (R2) | 113 (0.3) | 130 (0.3) | 40 (0.1) | 16 (0.1) | 6 (0.2) | 2 (0.1) | 2 (0.2) | 0 (0.0) | ||||
| Unknown | 226 (0.6) | 167 (0.4) | 80 (0.3) | 61 (0.2) | 11 (0.4) | 5 (0.2) | 26 (2.2) | 13 (1.2) | ||||
Chi-squared.
FIGURE 2.
Kaplan-Meier survival curves for prostate, cervical, colon, and rectal cancers with log-rank test for equality of survivors [Color figure can be viewed at wileyonlinelibrary.com]
Patients who underwent open surgery had higher rates of mortality at 30 days for colon (3.7% vs 1.7%; HR, 2.16; 95% CI, 1.93-2.42; P < .001) and rectal (2.1% vs 1.2%; HR, 1.67; 95% CI, 1.07-2.60; P = .025) cancers. Mortality rates following open surgery remained higher at 90 days for colon (5.7% vs 2.4%; HR, 2.12; 95% CI, 1.95-2.32; P < .001) and rectal cancer (3.6% vs 2.3%; HR, 1.62; 95% CI, 1.16-2.28; P = .005). There was no difference in 30-day or 90-day mortality, by surgical approach, for patients who underwent radical prostatectomy or radical hysterectomy (Table 3).
TABLE 3.
Comparison of adjusted survival and postoperative outcomes by disease site
| HR (95% CI) |
||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Prostate |
Colon |
Rectal |
Cervical |
|||||||||
| Open | MIS | P value | Open | MIS | P value | Open | MIS | P value | Open | MIS | P value | |
| 5-y Survival | 1.18 (1.05-1.33) | REF | .005 | 1.45 (1.39-1.51) | REF | <.001 | 1.28 (1.10-1.50) | REF | .002 | 0.61 (0.44-0.82) | REF | .003 |
| 30-d Mortality | 1.34 (0.92-1.94) | REF | .125 | 2.16 (1.93-2.42) | REF | <.001 | 1.67 (1.07-2.60) | REF | 0.025 | 0.47 (0.04-5.17) | REF | .537 |
| 90-d Mortality | 1.17 (0.85-1.59) | REF | .335 | 2.12 (1.95-2.32) | REF | <.001 | 1.62 (1.16-2.28) | REF | 0.005 | 1.41 (0.24-8.33) | REF | .703 |
| OR (95% CI) | OR (95% CI) | OR (95% CI) | OR (95% CI) | |||||||||
| Readmission at 30 d | 1.33 (0.93-1.91) | REF | .121 | 1.15 (1.05-1.26) | REF | .003 | 0.90 (0.73-1.09) | REF | 0.285 | 0.78 (0.50-1.22) | REF | .285 |
| Positive margins | 1.08 (0.97-1.21) | REF | .168 | 1.77 (1.56-2.02) | REF | <.001 | 1.31 (0.97-1.78) | REF | 0.074 | 1.01 (0.70-1.46) | REF | .960 |
3.3 |. Postoperative outcomes
Patients who underwent open surgery for colon cancer had higher rates of readmission at 30 days postoperatively (6.8% vs 5.9%; OR, 1.15; 95% CI, 1.05-1.26; P = .003). There was no difference in 30-day readmission between open surgery and MIS following resection of prostate, rectal, or cervical cancer. Patients who underwent open surgery for colon cancer had higher rates of positive margins (2.7% vs 1.6%; OR, 1.77; 95% CI, 1.56-2.02; P < .001). There was no difference in positive margin rate between open surgery and MIS following resection of prostate, rectal, or cervical cancer.
3.4 |. Sensitivity analyses
Similar to the primary results, inverse probability weighted analyses demonstrated that patients who underwent open surgery had worse 5-year survival for prostate cancer (HR, 1.25; 95% CI, 1.13-1.39; P < .001), colon cancer (HR, 1.38; 95% CI, 1.32-1.43; P < .001), and rectal cancer (HR, 1.27; 95% CI, 1.13-1.4; P < .001). Open surgery for cervical cancer was again associated with improved 5-year survival (HR, 0.68; 95% CI, 0.52-0.89; P = .005). Conditional survival analyses, measured as survival amongst those surviving beyond 90 days after surgery, produced results similar to the primary analyses. The 5-year survival remained worse following open surgery for prostate (HR, 1.17; 95% CI, 1.04-1.32; P = .008), colon (HR, 1.32; 95% CI, 1.26-1.38; P < .001), and rectal (HR, 1.19; 95% CI, 1.01-1.41; P = .036) cancers. Open surgery was again associated with improved 5-year survival for cervical (HR, 0.61; 95% CI, 0.44-0.85; P = .003) cancer.
Repeated survival analyses using MIS vs open surgery, with patients converted to open included in the open surgery cohort, yielded very similar results. The 5-year survival was again worse following open surgery for prostate (HR, 1.16; 95% CI, 1.03-1.30; P = .014), colon (HR, 1.37; 95% CI, 1.32-1.4; P < .001), and rectal (HR, 1.24; 95% CI, 1.09-1.41; P = .001) cancers. Open surgery continued to be associated with improved 5-year survival for cervical (HR, 0.63; 95% CI, 0.46-0.87; P = .005) cancer. Adjusting for the use of adjuvant therapy also yielded minimal change in results, with worse 5-year survival following open surgery for prostate (HR, 1.17; 95% CI, 1.04-1.31; P = .007), colon (HR, 1.46; 95% CI, 1.40-1.52; P < .001) and rectal (HR, 1.29; 95% CI, 1.11-1.51; P = .001) cancer, and improved 5-year survival following open surgery for cervical cancer (HR, 0.59; 95% CI, 0.43-0.81; P = .001).
4 |. DISCUSSION
In this study, a propensity-score-matched national cohort of patients undergoing surgical resection of early stage prostate, colon, rectal, or cervical cancer was analyzed to compare 5-year survival and postoperative outcomes based on surgical approach. The results suggest that open surgery for early stage prostate, colon, and rectal cancer is associated with worse 5-year survival when compared to those patients undergoing MIS. Open surgery for early stage cervical cancer is associated with improved 5-year survival as compared to MIS. Open rectal cancer surgery was less commonly associated with readmission to the hospital within 30 days of surgery, and open colon surgery was more commonly associated with positive surgical margins, however, there were no differences in rates of readmission or positive surgical margins for the other disease sites evaluated. These findings remained consistent despite the use of different analytic approaches and multiple sensitivity analyses.
Our findings use contemporary data to reinforce the literature that has established the precedent for adoption of MIS techniques for abdominopelvic cancer surgery. Despite known limitations, prior RCTs have demonstrated equivalent or superior outcomes following MIS for colon, rectal, and prostate cancer surgery.8–11,15 Non-randomized, center-specific cohorts and large database studies from the early MIS era have also returned similar findings.12,13,39–42 However, especially in light of the recent findings related to cervical cancer, the present analysis importantly uses more contemporary data to confirm these findings. This investigation argues that despite broad use of MIS in more diverse practice environments there has not been a demonstrable decrement in survival for patients with prostate, colon, and rectal cancer.
The confirmation and validation of prior reports regarding cervical cancer outcomes after MIS are important for several reasons. First, this reinforces concerns that MIS may not be best suited for the treatment of cervical cancer. These prior studies have led to recent modifications in clinical guidelines from organizations at the forefront of cervical cancer care who now caution against the use of MIS hysterectomy for this disease.43,44 Our investigation further validates that those statements appear warranted. However, as previously mentioned, the cervical cancer literature has led to increased regulatory and media scrutiny regarding MIS cancer surgery more broadly. In this manuscript, we use similar methods, the same data set, and have similar findings regarding cervical cancer surgery as one of the two studies that has led to recent attention to this subject. Yet, in applying those methods to evaluate for a broader issue with MIS cancer surgery across other cancer sites, we do not find an association between MIS and diminished survival for patients with early stage colon, rectal, or prostate cancer. While the exact cause of inferior survival associated with MIS for early stage cervical cancer remains unclear, one proposed mechanism is the potential for peritoneal dissemination of tumor cells caused by contamination from the vagina at the time of colpotomy or transvaginal specimen removal. This may reconcile the observed differences in survival, as the tumor is encased within tissue during the entire resection of early stage prostate, colon, or rectal cancers.
This study should be interpreted within the context of several limitations. First, the retrospective design of the study introduces potential selection bias and limits the results to association rather than causation. Thus, we caution against using this data as an inference to the superiority of any surgical approach where an association between technique and survival has been demonstrated. Although such conclusions should be avoided, this investigation remains valuable for its lack of detection of an MIS related survival decrease for prostate, colon, and rectal cancer surgery. Second, there is a lack of data regarding disease recurrence and cause of death in the NCDB. These, along with other potential unmeasured patient characteristics, may have influenced the results of this study. Third, the follow up duration of this study, due to desire to use contemporary data, was limited to 5 years, and this may be a shorter window than is required to pick up more substantial differences in survival for these early stage cancers. Although this is worth consideration, it remains that the previous reports from the cervical cancer literature both using NCDB and in a randomized controlled trial demonstrated rather rapid divergence of survival curves attributable to surgical approach. Thus, our intent was to evaluate for a similar impact at other sites. In addition, lymph node harvest represents a key component for resection of colon and rectal malignancies, however given the proportion of cases missing information regarding the number of lymph nodes harvested we excluded this analysis from the current work. While this may potentially bias results, it would likely bias towards improved survival for open resection, which is contrary to the reported findings.
These limitations notwithstanding, the results of this study suggest that the recently reported findings of decreased survival following MIS resection of early stage cervical cancer, while valid, do not appear to extend to colon, rectal, and prostate cancer surgery. We demonstrate a survival advantage associated with MIS resection of colon, rectal, and prostate cancers, however, that finding should be interpreted with caution before making inferences about open surgery in a study that by design does not evaluate for a causal influence. This finding may indicate a need for larger and more contemporary randomized surgical trials at these disease sites.
Given these findings, we do not demonstrate a necessity to expand recent warnings regarding the use of MIS techniques for cervical cancer to the other abdominopelvic malignancies evaluated. In the context of these and other recent findings, ongoing and continual evaluation of outcomes should be strongly encouraged before adoption of new surgical techniques for the management of cancer. Findings for one disease site may not translate to similar findings at other sites underscoring the necessity for disease specific research and regulatory guidance.
5 |. CONCLUSION
Open surgery was associated with worse 5-year survival following resection of early stage colon, rectal, and prostate cancer, and was again associated with improved survival following resection of early stage cervical cancer. This suggests that MIS is an acceptable alternative to open surgery in selected patients for early stage prostate, colon, and rectal cancer, while recent concerns regarding MIS for cervical cancer remain.
Supplementary Material
ACKNOWLEDGMENTS
This study was supported by the Northwestern Institute for Comparative Effectiveness Research in Oncology (NICER-Onc) of the Robert H. Lurie Comprehensive Cancer Center at the Feinberg School of Medicine, Northwestern University. TKY and RJE (Agency for Healthcare Research and Quality [AHRQ] 5T32HS000078) were supported by a postdoctoral research fellowship. ADY’s research is supported by the National Heart Lung and Blood Institute of the National Institutes of Health (K08HL145139).
Footnotes
CONFLICT OF INTERESTS
The authors declare that there are no conflict of interests.
DATA AVAILABILITY STATEMENT
Data sharing is not applicable to this article as no new data were created or analyzed in this study. The data that support the findings of this study are available from The American College of Surgeons. Restrictions apply to the availability of these data, which were used under license for this study.
SUPPORTING INFORMATION
Additional supporting information may be found online in the Supporting Information section.
REFERENCES
- 1.Siddaiah-Subramanya M, Tiang KW, Nyandowe M. A new era of minimally invasive surgery: progress and development of major technical innovations in general surgery over the last decade. Surg J. 2017;3(4):e163–e166. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Tsui C, Klein R, Garabrant M. Minimally invasive surgery: national trends in adoption and future directions for hospital strategy. Surg Endosc. 2013;27(7):2253–2257. [DOI] [PubMed] [Google Scholar]
- 3.Fantus RJ, Cohen A, Riedinger CB, et al. Facility-level analysis of robot utilization across disciplines in the National Cancer Database. J Robot Surg. 2019;13(2):293–299. [DOI] [PubMed] [Google Scholar]
- 4.Wichmann MW, Meyer G, Angele MK, Schildberg FW, Rau HG. Recent advances in minimally invasive colorectal cancer surgery. Onkologie. 2002;25(4):318–323. [DOI] [PubMed] [Google Scholar]
- 5.Banegas MP, Harlan LC, Mann B, Yabroff KR. Toward greater adoption of minimally invasive and nephron-sparing surgical techniques for renal cell cancer in the United States. Urol Oncol. 2016; 34(10):433 e439–433 e417. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Selby LV, DeMatteo RP, Tholey RM, et al. Evolving application of minimally invasive cancer operations at a tertiary cancer center. J Surg Oncol. 2017;115(4):365–370. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Sun Z, Kim J, Adam MA, et al. Minimally invasive versus open low anterior resection: equivalent survival in a national analysis of 14,033 patients with rectal cancer. Ann Surg. 2016;263(6):1152–1158. [DOI] [PubMed] [Google Scholar]
- 8.Clinical Outcomes of Surgical Therapy Study G, Nelson H, Sargent DJ, et al. A comparison of laparoscopically assisted and open colectomy for colon cancer. N Engl J Med. 2004;350(20):2050–2059. [DOI] [PubMed] [Google Scholar]
- 9.Fleshman J, Sargent DJ, Green E, et al. Laparoscopic colectomy for cancer is not inferior to open surgery based on 5-year data from the COST Study Group trial. Ann Surg. 2007;246(4):655–662. discussion 662-654. [DOI] [PubMed] [Google Scholar]
- 10.Lacy AM, Garcia-Valdecasas JC, Delgado S, et al. Laparoscopy-assisted colectomy versus open colectomy for treatment of non-metastatic colon cancer: a randomised trial. Lancet. 2002;359(9325):2224–2229. [DOI] [PubMed] [Google Scholar]
- 11.Yaxley JW, Coughlin GD, Chambers SK, et al. Robot-assisted laparoscopic prostatectomy versus open radical retropubic prostatectomy: early outcomes from a randomised controlled phase 3 study. Lancet. 2016;388(10049):1057–1066. [DOI] [PubMed] [Google Scholar]
- 12.Bilimoria KY. Use and outcomes of laparoscopic-assisted colectomy for cancer in the United States. Arch Surg. 2008;143(9):832–839. discussion 839-840. [DOI] [PubMed] [Google Scholar]
- 13.Hu JC. Comparative effectiveness of minimally invasive vs open radical prostatectomy. JAMA. 2009;302(14):1557–1564. [DOI] [PubMed] [Google Scholar]
- 14.Ulmer WD, Prasad SM, Kowalczyk KJ, et al. Factors associated with the adoption of minimally invasive radical prostatectomy in the United States. J Urol. 2012;188(3):775–780. [DOI] [PubMed] [Google Scholar]
- 15.Bonjer HJ, Deijen CL, Abis GA, et al. A randomized trial of laparoscopic versus open surgery for rectal cancer. N Engl J Med. 2015;372(14):1324–1332. [DOI] [PubMed] [Google Scholar]
- 16.Birkmeyer JD, Sun Y, Wong SL, Stukel TA. Hospital volume and late survival after cancer surgery. Ann Surg. 2007;245(5):777–783. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Sharma V, Karnes RJ. Much ado about robotic versus open radical prostatectomy. Lancet Oncol. 2018;19(8):1003–1004. [DOI] [PubMed] [Google Scholar]
- 18.Parsons JK, Messer K, Palazzi K, Stroup SP, Chang D. Diffusion of surgical innovations, patient safety, and minimally invasive radical prostatectomy. JAMA Surg. 2014;149(8):845–851. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Barkun JS, Aronson JK, Feldman LS, Maddern GJ, Strasberg SM. Evaluation and stages of surgical innovations. Lancet. 2009; 374(9695):1089–1096. [DOI] [PubMed] [Google Scholar]
- 20.Wilson CB. Adoption of new surgical technology. BMJ. 2006; 332(7533):112–114. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Melamed A, Margul DJ, Chen L, et al. Survival after minimally invasive radical hysterectomy for early-stage cervical cancer. N Engl J Med. 2018;379(20):1905–1914. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Ramirez PT, Frumovitz M, Pareja R, et al. Minimally invasive versus abdominal radical hysterectomy for cervical cancer. N Engl J Med. 2018;379(20):1895–1904. [DOI] [PubMed] [Google Scholar]
- 23.Caution when using robotically-assisted surgical devices in women’s health including mastectomy and other cancer-related surgeries: FDA safety communication. https://www.fda.gov/MedicalDevices/Safety/AlertsandNotices/ucm632142.htm Accessed March 20, 2019.
- 24.Bilimoria KY, Stewart AK, Winchester DP, Ko CY. The National Cancer Data Base: a powerful initiative to improve cancer care in the United States. Ann Surg Oncol. 2008;15(3):683–690. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Boffa DJ, Rosen JE, Mallin K, et al. Using the national cancer database for outcomes research: a review. JAMA Oncol. 2017;3(12):1722–1728. [DOI] [PubMed] [Google Scholar]
- 26.American College of Surgeons. About the National Cancer Database. Available at: https://www.facs.org/quality-programs/cancer/ncdb/about
- 27.World Health Organization. International Classification of Diseases for Oncology, 3rd ed Available at: https://www.who.int/classifications/icd/adaptations/oncology/en/
- 28.United States Department of Agriculture Economic Research Service. Rural-urban continuum codes. [Google Scholar]
- 29.Klabunde CN, Potosky AL, Legler JM, Warren JL. Development of a comorbidity index using physician claims data. J Clin Epidemiol. 2000; 53(12):1258–1267. [DOI] [PubMed] [Google Scholar]
- 30.Commission on Cancer. About cancer program categories. Available at: https://www.facs.org/quality-programs/cancer/coc/apply/categories
- 31.Pecorelli S, Zigliani L, Odicino F. Revised FIGO staging for carcinoma of the cervix. Int J Gynaecol Obstet. 2009;105(2):107–108. [DOI] [PubMed] [Google Scholar]
- 32.Austin PC. An introduction to propensity score methods for reducing the effects of confounding in observational studies. Multivariate Behav Res. 2011;46(3):399–424. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Garrido MM, Kelley AS, Paris J, et al. Methods for constructing and assessing propensity scores. Health Serv Res. 2014;49(5):1701–1720. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Rosenbaum PR. Model-based direct adjustment. J Am Stat Assoc. 1987;82(398):387–394. [Google Scholar]
- 35.Austin PC. The use of propensity score methods with survival or time-to-event outcomes: reporting measures of effect similar to those used in randomized experiments. Stat Med. 2014;33(7):1242–1258. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Austin PC, Stuart EA. Moving towards best practice when using inverse probability of treatment weighting (IPTW) using the propensity score to estimate causal treatment effects in observational studies. Stat Med. 2015;34(28):3661–3679. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Lunceford JK, Davidian M. Stratification and weighting via the propensity score in estimation of causal treatment effects: a comparative study. Stat Med. 2004;23(19):2937–2960. [DOI] [PubMed] [Google Scholar]
- 38.Hieke S, Kleber M, Konig C, Engelhardt M, Schumacher M. Conditional Survival: A useful concept to provide information on how prognosis evolves over time. Clin Cancer Res. 2015;21(7):1530–1536. [DOI] [PubMed] [Google Scholar]
- 39.Bonjer HJ, Hop WC, Nelson H, et al. Laparoscopically assisted vs open colectomy for colon cancer: a meta-analysis. Arch Surg. 2007; 142(3):298–303. [DOI] [PubMed] [Google Scholar]
- 40.Fan CJ, Chien HL, Weiss MJ, et al. Minimally invasive versus open surgery in the Medicare population: a comparison of post-operative and economic outcomes. Surg Endosc. 2018;32(9):3874–3880. [DOI] [PubMed] [Google Scholar]
- 41.Hyung WJ, Cheong JH, Kim J, Chen J, Choi SH, Noh SH. Application of minimally invasive treatment for early gastric cancer. J Surg Oncol. 2004;85(4):181–185. [DOI] [PubMed] [Google Scholar]
- 42.Vennix S, Pelzers L, Bouvy N, et al. Laparoscopic versus open total mesorectal excision for rectal cancer. Cochrane Database Syst Rev. 2014; (4), 10.1002/14651858.CD005200.pub3 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.National Comprehensive Cancer Network (NCCN). Cervical Cancer (Version 3.2019). Available at: https://www.nccn.org/professionals/physician_gls/PDF/cervical.pdf Accessed March 20, 2019.
- 44.Statement from the Society of Gynecologic Oncology. Availale at: https://www.sgo.org/clinical-practice/guidelines/notice-to-sgo-members-emerging-data-on-the-surgical-approach-for-radical-hysterectomy-in-the-treatment-of-women-with-cervical-cancer/ Accessed April 15, 2019.
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.