Abstract
Background:
Cervical cancer incidence in premenopausal women is rising, and fertility-sparing surgery offers an important option in this young population. There is a lack of evidence about what tumor size cutoff should be used to define candidacy for fertility-sparing surgery.
Objectives:
We sought to describe how the association between fertility-sparing surgery (compared with standard surgery) and life expectancy varies by tumor size among patients with cervical cancers measuring ≤ 4 cm in largest diameter. Our secondary objective was to quantify the probability of undergoing adjuvant radiotherapy among patients who underwent fertility-sparing surgery as a function of tumor size.
Study design:
We identified patients in the National Cancer Database aged ≤45 years, diagnosed with stage I cervical cancer with tumors ≤4 cm between 2006 and 2018, who received no preoperative radiation or chemotherapy and underwent either fertility-sparing surgery (cone or trachelectomy, either simple or radical) or standard surgery (simple or radical hysterectomy) as their primary treatment. Propensity-score matching was performed to compare patients who underwent fertility-sparing surgery with those who received standard surgery. A flexible parametric model was employed to quantify the difference in life expectancy within 5 years of diagnosis (restricted mean survival time; RMST) based on tumor size for patients who underwent fertility-sparing versus standard surgery. Additionally, among those who underwent fertility-sparing surgery, a logistic regression model was used to explore the relationship between tumor size and the probability of receiving adjuvant radiation.
Results:
A total of 11,946 patients met the inclusion criteria, of whom 904 (7.6%) underwent fertility-sparing surgery. After propensity-score matching, 897 patients who underwent fertility-sparing surgery were matched 1:1 with those who underwent standard surgery. While 5-year life expectancy was similar among patients who had fertility sparing or standard surgery regardless of tumor sizes, estimates of life-expectancy differences associated with fertility-sparing surgery were more precise among patients with smaller tumors (1-cm tumor: RMST difference, −0.10 months; 95% CI, −0.67 to 0.47) compared with those with larger tumors (4-cm tumor: RMST difference, −0.11 months; 95% CI, −3.79 to 3.57). The probability of receiving adjuvant radiation increased with tumor size, ranging from 5.6% (95% CI, 3.9–7.9%) for a 1-cm tumor to 37% (95% CI, 24.3–51.8%) for a 4-cm tumor.
Conclusion:
Within 5 years of diagnosis, young patients with stage I cancers measuring ≤4 cm had similar survival outcomes following either fertility-sparing or standard surgery. However, because few patients with tumors >2 cm underwent fertility-sparing surgery, a clinically important survival difference could not be excluded in this population.
Keywords: early-stage cervical cancer, fertility-sparing, survival, flexible parametric restricted mean survival time, survival model
Tweetable statement:
Five-year life expectancy among early-stage cervical cancer survivors with tumors ≤ 4 cm did not differ between those who received fertility-sparing versus standard surgery, irrespective of tumor size.
Introduction
The incidence of cervical cancer in the premenopausal population is increasing.1 Among women younger than 45 years, cervical cancer ranks in the top three cancers in 80% of countries worldwide2 and affects many women who have not completed childbearing. Understanding who may benefit from fertility-sparing treatment without compromising oncologic outcomes is vital for patients and providers to make informed decisions. Prior studies have demonstrated that offering fertility-sparing options and counseling improves patient quality of life3 and ability to cope with a cancer diagnosis.4
International guidelines offer tumor-size-specific recommendations regarding fertility-sparing procedures for cervical cancer. National Comprehensive Cancer Network (NCCN) guidelines5 offer fertility-sparing procedures as primary treatment for patients with tumors ≤2 cm in diameter. However, the NCCN guidelines caution against fertility-sparing procedures for those with tumors measuring 2–4 cm due to less validation and a higher risk of requiring adjuvant therapy. European guidelines classify fertility-sparing treatments in patients with tumors >2 cm as experimental and thus not recommended as standard treatment.6 However, despite its wide acceptance, a 2-cm cutoff is arbitrary, has not been rigorously compared with alternative thresholds, and appears to be derived from studies that investigated the risk of disease recurrence7 rather than the appropriateness of the fertility-sparing procedure itself. The current guidelines rely on a limited number of studies, some of which excluded patients with tumors larger than 2 cm.8–11 The paucity of evidence has led to inconclusive findings, with some studies suggesting that proceeding with fertility-sparing procedures for tumors of up to 4 cm may be reasonable,12–15 while others suggest increased oncologic risk above the 2 cm threshold.16 Thus, the appropriate size threshold above which fertility-sparing surgery is inadvisable remains controversial.
In this study, we used a large national cancer registry to generate evidence about how tumor size relates to the safety of fertility-sparing surgery among young patients with stage I cervical cancer. To this end, we quantified whether there exists a size cutoff above which fertility-sparing surgery is associated with inferior oncologic outcomes than standard surgery. Furthermore, among patients who received fertility-sparing surgery, we determined the association between tumor size and receipt of postoperative radiotherapy, to understand whether the effort to preserve fertility was futile among patients with larger tumors.
Materials and Methods
Data source
We used the most recent available file with complete survival data in the National Cancer Database (NCDB) to perform a retrospective cohort study. This database, a joint project of the Commission on Cancer and the American Cancer Society, is sourced from hospital registry data collected from more than 1,500 facilities. It represents more than 70% of all newly diagnosed cases of cancer in the United States and more than 85% of newly diagnosed cervical cancer cases.17 The data are deidentified; therefore, exemption was granted by the Institutional Review Board at MD Anderson Cancer Center. The Commission on Cancer and the hospitals submitting data to the NCDB have not verified and are not responsible for the statistical validity of the data analysis or the conclusions of this study.
Cohort selection
Using the 2019 NCDB public use file, we identified patients ≤45 years old who were diagnosed with cervical cancer from 2006 to 2018 (Figure 1). Age cutoff was utilized due to natural decline in fertility18 and to be consistent with prior studies19. We included patients who underwent either fertility-sparing surgery (cone or trachelectomy, simple or radical) or standard surgery (hysterectomy, simple or radical) as their primary treatment, irrespective of the surgical approach (minimally invasive or open), for cancers staged as IA1 to IB2 with the American Joint Commission on Cancer (AJCC) staging system. The AJCC staging criteria were based on both the 7th and 8th editions of the AJCC Cancer Staging Manual20,21 and were aligned with the International Federation of Gynecology and Obstetrics (FIGO) 1994 and 2009 staging criteria. The clinical stage was used for staging when the pathologic stage was not available or clinical stage was the first stage entered in the database. Patients with squamous cell carcinoma, adenocarcinoma, or adenosquamous carcinoma histology confirmed on microscopy were included, using codes from the International Classification of Diseases for Oncology, 3rd Edition (Box 1).22 We excluded patients with a prior cancer diagnosis, primary radiation therapy, primary or unknown chemotherapy status, or tumors >4 cm or of unknown size. For patients with stage IA2 or greater, we excluded patients who did not undergo pelvic lymph node assessment and those for whom the lymph node status was unknown. Patients who underwent only a loop electrocautery excision procedure (LEEP) or biopsy or whose type of surgery was unknown were excluded. We excluded LEEPs because cone biopsy is the preferred excisional procedure in the United States and because it was not feasible to determine whether LEEPs were intended as a cancer-directed surgery or for other purposes.
Figure 1:
Selection of cohort
Exposure and outcome
The exposures of interest were the type of surgery (fertility-sparing or standard) and tumor size. We considered patients who underwent conization or trachelectomy (simple or radical) to have had fertility-sparing surgery and those who underwent hysterectomy (simple or radical) to have had standard surgery.
The primary outcome was the difference in life expectancy within 5 years of diagnosis between patients who underwent fertility-sparing surgery and those who received standard surgery. The secondary outcome was the probability of receiving adjuvant radiotherapy following fertility-sparing surgery to assess for futility of the initial fertility-sparing procedure from a fertility perspective. The primary and secondary outcomes were evaluated across tumor size.
Differences in 5-year life expectancy were quantified using restricted mean survival time (RMST).23,24 This measure of survival is calculated by estimating the area under a survival curve between the origin time (time of cancer diagnosis) and 5 years after diagnosis. The difference in 5-year RMST quantifies the gain or loss in life expectancy within 5 years of diagnosis that is associated with undergoing fertility-sparing surgery compared with standard surgery. Difference in RMST is a well-characterized measure used to compare survival between groups and is recommended as an endpoint in interventional and observational studies due to its interpretability and efficiency.25 In contrast to the more commonly used hazard ratio (usually calculated from a Cox model), the difference in life expectancy is meaningful even when the proportional hazards assumption is violated. Furthermore, it is difficult to interpret the clinical significance of differing hazard ratios when they are obtained from populations with large differences in absolute risk of death, such as patients with small versus large cervical tumors. Conversely, difference in 5-year life expectancy (i.e., RMST) is an absolute measure of survival difference, which has the same clinical significance regardless of absolute risk. A final advantage of this approach is that estimates of RMST utilize all information in the survival curve, making them more precise and informative than estimates of survival at a given time point, such as 5-year survival.
Propensity score matching
Propensity-score matching was used to create a cohort in which the patients who underwent fertility-sparing surgery and the patients who underwent standard surgery were balanced on covariates previously found to affect likelihood of fertility-sparing surgery, which could confound the association between the treatment and life expectancy.26,27,28 Each patient who underwent fertility-sparing surgery was matched 1:1 with a patient who did not undergo fertility-sparing surgery using Greedy nearest-neighbor matching. Standardized mean differences of the covariates in the propensity-matched cohort were assessed for balance.
Covariates included in the propensity-score matching included tumor, patient, and treatment center characteristics. Tumor characteristics were: size (measured in the largest dimension in millimeters), grade (1, 2, 3, or unknown), and histology (squamous cell carcinoma, adenocarcinoma, adenosquamous carcinoma). Patient and treatment center characteristics were: age at diagnosis (continuous, in years), race and ethnicity (American Indian, Asian, non-Hispanic Black, Hispanic, non-Hispanic White, or unknown), year of diagnosis, Charlson Comorbidity Index (0–3), annual median household income for the ZIP code of patient’s residence (quintiles), insurance status (private, Medicaid or other government insurance, Medicare, none, unknown), rural/urban status based on state and county FIPS code (metropolitan, urban, rural, or unknown), and educational status (percentage of residents with no high school diploma in the ZIP code of residence). Race and ethnicity data were used in this study as prior studies demonstrated differences in fertility-sparing procedures, incidence, and survival outcomes by race and ethnicity.26,29
Statistical analyses
An assessment of overall survival was performed using the Kaplan-Meier method in the propensity score-matched cohort. Next, to quantify how the association between fertility-sparing surgery and survival varied by tumor size, we fit a flexible parametric model in the propensity score matched cohort that included an indicator variable for fertility-sparing surgery, tumor size (modeled as a restricted cubic spline with 3 knots), and interaction terms between these variables. This model makes no assumption about the shape of the relationship between the survival function and tumor size and is convenient for estimating differences in RMST as a function of size.25,30 We plotted the predicted differences in 5-year RMST and inspected this plot for evidence of a nonlinear relationship or inflection points.
We hypothesized that for patients with larger tumors, 5-year RMST would be lower among those who underwent fertility-sparing surgery compared with those who underwent traditional surgery. If the predicted 95% confidence band for the difference in 5-year RMST excluded 0 months for any area of the tumor size RMST curve, we would reject the null hypothesis of no difference. Because we planned to use all available data, and since there is no closed-form power/sample size calculation for this study design, we did not perform a power calculation a priori. Instead, we relied on the range of the 95% confidence intervals to understand whether failure to reject the null hypothesis also provided strong evidence of the absence of a clinically significant effect. We considered a 3-month difference in 5-year RSMT to be clinically significant.
Among patients who underwent fertility-sparing surgery, we fit a logistic regression model with receipt of postoperative radiation as the dependent variable and tumor size (modeled as a restricted cubic spline with 3 knots) as the independent variable. This model was used to estimate the probability of postoperative radiotherapy by tumor size with associated 95% confidence intervals (CIs)
Sensitivity analyses were performed to assess the robustness of the primary results to model specifications. The flexible parametric model was refitted using 4 and 5 knots. All calculations were performed using SAS Enterprise Guide 7.1 software. A P-value of <0.05 was considered significant.
Results
A total of 11,946 patients met the inclusion criteria, of whom 904 (7.6%) underwent fertility-sparing surgery and 11,042 underwent standard surgery (Figure 1). Of those who underwent fertility-sparing surgery, 503 (55.6%) had a cone procedure and 401 (44.3%) had a trachelectomy. Compared to patients who underwent standard surgery, patients who underwent fertility-sparing surgery were younger, had smaller tumors, and had fewer comorbidities (Table 1). Patients who underwent fertility-sparing surgery were more often non-Hispanic White or Asian; residing in more educated, wealthier, and metropolitan ZIP codes; and less often uninsured or insured by Medicaid or another government program.
Table 1:
Characteristics of the study patients
| Variable* | Before matching | After matching | ||||
|---|---|---|---|---|---|---|
| Fertility-sparing N=904 |
Standard surgery N=11,042 |
P value | Fertility-sparing N=897 |
Standard surgery N=897 |
P value | |
| Age at diagnosis (mean, SD) | 32.27 (5.39) | 37.05 (5.30) | <.001 | 32.3 (5.36) | 32.5 (5.36) | 0.495 |
| Race/ethnicity | 0.012 | 0.99 | ||||
| Non-Hispanic White | 643 (71.1%) | 7688 (69.6%) | 641 (71.5%) | 644 (71.8%) | ||
| Non-Hispanic Black | 69 (7.6%) | 1007 (9.1%) | 68 (7.6%) | 72 (8.0%) | ||
| American Indian | † | 68 (0.6%) | † | † | ||
| Asian | 51 (5.6%) | 408 (3.7%) | 48 (5.4%) | 42 (4.7%) | ||
| Hispanic | 120 (13.3%) | 1662 (15.1%) | 120 (13.4%) | 118 (13.2%) | ||
| Unknown | 19 (2.1%) | 209 (1.9%) | 18 (2.0%) | 19 (2.1%) | ||
| Tumor size | <.001 | 1 | ||||
| 0–5 mm | 377 (41.7%) | 2494 (22.6%) | 374 (41.7%) | 374 (41.7%) | ||
| 6–10 mm | 180 (19.9%) | 2062 (18.7%) | 180 (20.1%) | 180 (20.1%) | ||
| 11–15 mm | 114 (12.6%) | 1654 (15.0%) | 114 (12.7%) | 114 (12.7%) | ||
| 16–20 mm | 77 (8.5%) | 1261 (11.4%) | 76 (8.5%) | 76 (8.5%) | ||
| 21–25 mm | 57 (6.3%) | 1137 (10.3%) | 56 (6.2%) | 56 (6.2%) | ||
| 26–30 mm | 51 (5.6%) | 1022 (9.3%) | 49 (5.5%) | 49 (5.5%) | ||
| 31–35 mm | 22 (2.4%) | 763 (6.9%) | 22 (2.5%) | 22 (2.5%) | ||
| 36–40 mm | 26 (2.9%) | 649 (5.9%) | 26 (2.9%) | 26 (2.9%) | ||
| Year of diagnosis | <.001 | 0.931 | ||||
| 2006 | 32 (3.5%) | 656 (5.9%) | 32 (3.6%) | 32 (3.6%) | ||
| 2007 | 36 (4.0%) | 665 (6.0%) | 36 (4.0%) | 41 (4.6%) | ||
| 2008 | 38 (4.2%) | 647 (5.9%) | 38 (4.2%) | 34 (3.8%) | ||
| 2009 | 47 (5.2%) | 811 (7.3%) | 47 (5.2%) | 43 (4.8%) | ||
| 2010 | 59 (6.5%) | 828 (7.5%) | 59 (6.6%) | 46 (5.1%) | ||
| 2011 | 69 (7.6%) | 822 (7.4%) | 69 (7.7%) | 60 (6.7%) | ||
| 2012 | 78 (8.6%) | 846 (7.7%) | 78 (8.7%) | 79 (8.8%) | ||
| 2013 | 68 (7.5%) | 898 (8.1%) | 67 (7.5%) | 72 (8.0%) | ||
| 2014 | 89 (9.8%) | 930 (8.4%) | 89 (9.9%) | 89 (9.9%) | ||
| 2015 | 105 (11.6%) | 950 (8.6%) | 103 (11.5%) | 125 (13.9%) | ||
| 2016 | 104 (11.5%) | 1023 (9.3%) | 102 (11.4%) | 106 (11.8%) | ||
| 2017 | 101 (11.2%) | 964 (8.7%) | 99 (11.0%) | 99 (11.0%) | ||
| 2018 | 78 (8.6%) | 1002 (9.1%) | 78 (8.7%) | 71 (7.9%) | ||
| Charlson Comorbidity Index | 0.002 | 0.923 | ||||
| 0 | 851 (94.1%) | 10,027 (90.8%) | 844 (94.1%) | 840 (93.6%) | ||
| 1 | 51 (5.6%) | 855 (7.7%) | 51 (5.7%) | 55 (6.1%) | ||
| 2 | 2 (0.2%) | 111 (1.0%) | 2 (0.2%) | 2 (0.2%) | ||
| 3+ | 0 (0.0%) | 49 (0.4%) | 0 (0.0%) | 0 (0.0%) | ||
| Median annual household income in ZIP code | <.001 | 0.939 | ||||
| <$40,227 | 113 (12.5%) | 1927 (17.5%) | 113 (12.6%) | 109 (12.2%) | ||
| $40,227-$50,353 | 159 (17.6%) | 2269 (20.5%) | 158 (17.6%) | 168 (18.7%) | ||
| $50,354-$63,332 | 203 (22.5%) | 2389 (21.6%) | 201 (22.4%) | 195 (21.7%) | ||
| ≥$63,333 | 336 (37.2%) | 3209 (29.1%) | 332 (37.0%) | 325 (36.2%) | ||
| Unknown | 93 (10.3%) | 1248 (11.3%) | 93 (10.4%) | 100 (11.1%) | ||
| Insurance | <.001 | 0.995 | ||||
| Private | 679 (75.1%) | 7384 (66.9%) | 674 (75.1%) | 682 (76.0%) | ||
| Medicaid or other government program | 169 (18.7%) | 2539 (23.0%) | 167 (18.6%) | 161 (17.9%) | ||
| Medicare | 7 (0.8%) | 257 (2.3%) | 7 (0.8%) | 7 (0.8%) | ||
| None | 37 (4.1%) | 704 (6.4%) | 37 (4.1%) | 35 (3.9%) | ||
| Unknown | 12 (1.3%) | 158 (1.4%) | 12 (1.3%) | 12 (1.3%) | ||
| Rurality | 0.002 | 0.974 | ||||
| Metropolitan | 775 (85.7%) | 9091 (82.3%) | 769 (85.7%) | 769 (85.7%) | ||
| Urban | 77 (8.5%) | 1396 (12.6%) | 77 (8.6%) | 78 (8.7%) | ||
| Rural | 11 (1.2%) | 158 (1.4%) | 11 (1.2%) | 9 (1.0%) | ||
| Unknown | 41 (4.5%) | 397 (3.6%) | 40 (4.5%) | 41 (4.6%) | ||
| Percentage of ZIP code residents without high school diploma | <.001 | 0.971 | ||||
| ≥17.6% | 167 (18.5%) | 2478 (22.4%) | 166 (18.5%) | 159 (17.7%) | ||
| 10.9%−17.5% | 204 (22.6%) | 2611 (23.6%) | 201 (22.4%) | 204 (22.7%) | ||
| 6.3%−10.8% | 213 (23.6%) | 2645 (24.0%) | 213 (23.7%) | 216 (24.1%) | ||
| <6.3% | 227 (25.1%) | 2076 (18.8%) | 224 (25.0%) | 218 (24.3%) | ||
| Unknown | 93 (10.3%) | 1232 (11.2%) | 93 (10.4%) | 100 (11.1%) | ||
| Grade | <.001 | 0.99 | ||||
| 1 | 205 (22.7%) | 2417 (21.9%) | 205 (22.9%) | 209 (23.3%) | ||
| 2 | 353 (39.0%) | 4810 (43.6%) | 351 (39.1%) | 346 (38.6%) | ||
| 3 | 136 (15.0%) | 2463 (22.3%) | 133 (14.8%) | 136 (15.2%) | ||
| Unknown | 210 (23.2%) | 1352 (12.2%) | 208 (23.2%) | 206 (23.0%) | ||
| Histology | 0.203 | 0.965 | ||||
| Adenocarcinoma | 368 (40.7%) | 4221 (38.2%) | 366 (40.8%) | 363 (40.5%) | ||
| Adenosquamous carcinoma | 34 (3.8%) | 515 (4.7%) | 34 (3.8%) | 36 (4.0%) | ||
| Squamous cell carcinoma | 502 (55.5%) | 6306 (57.1%) | 497 (55.4%) | 498 (55.5%) | ||
Numbers are n (%) unless otherwise specified.
Counts less than 10 are masked according to database requirements
After the propensity-score matching, 897 of the patients who underwent fertility-sparing surgery were successfully matched 1:1 with patients who did not undergo fertility-sparing surgery. After matching, patient and tumor characteristics were no longer significantly different between patients who underwent fertility-sparing versus standard surgery, and covariates were well balanced based on absolute standardized difference (Table 1 and Appendix A).
No difference was found in the overall survival between those who received fertility-sparing and standard surgery (Figure 2, p=0.77). Figure 3 shows the difference in 5-year life expectancy between individuals who underwent fertility-sparing surgery and propensity score–matched controls who underwent standard surgery. For all tumor sizes, the estimated differences in 5-year life expectancy between patients who underwent fertility-sparing and standard surgery ranged from −0.19 months to 0.19 months, and the 95% CIs included zero, suggesting that, compared with standard surgery, fertility-sparing surgery was not significantly associated with a difference in 5-year life expectancy for any tumor ≤4 cm. Similarly, there was no inflection point at 2 cm or at any other tumor size along the predicted line.
Figure 2: Overall survival of propensity score-matched cohort.
Solid black line represents the survival curve of those who underwent fertility-sparing surgery and gray line represents those who underwent standard surgery.
Figure 3: Difference in 5-year life expectancy between individuals who underwent fertility-sparing surgery and propensity score–matched controls who underwent standard surgery.
Circles represent actual data, with the center of each circle reflecting the average tumor size for each 5-mm category, and the sizes of the circles representing the number of patients included in the analysis. The solid line represents the predicted difference in 5-year life expectancy in months. The dashed lines represent 95% confidence intervals. The inset table represents restricted mean survival time (RMST) and 95% confidence interval by tumor size. A 5-year RMST difference >1 is in favor of fertility-sparing surgery, while <1 favors standard surgery.
Importantly, the 95% CIs for difference in 5-year life expectancy were narrower for smaller tumor sizes than for larger ones, limiting the precision of these estimates among patients with larger tumors (Figure 2). In a patient with a 1-cm tumor, the fertility-sparing surgery was associated with no more than a 0.67-month decrement in 5-year life expectancy (5-year RMST, −0.10; 95% CI, −0.67 to 0.47). In contrast, in a patient with a 4-cm tumor, the data were compatible with a potential 3.79-month decrement in 5-year life expectancy (5-year RMST, −0.11; 95% CI, −3.79 to 3.57).
Figure 4 plots the probability of receiving radiation following fertility-sparing surgery as a function of tumor size. As tumor size increased, the probability of radiation also increased, though the relationship was not linear. For a 1-cm tumor, the probability of receiving adjuvant radiation was 5.6% (95% CI, 3.9–7.9%), while for a 4-cm tumor the probability was 37% (95% CI, 24.3% to 51.8%).
Figure 4: Predicted probability of receiving adjuvant radiation following fertility-sparing surgery by tumor size.
Circles represent actual data, with the center of circle reflecting the average tumor size for each 5-mm category, and the sizes of the circles representing the number of patients included in the analysis. The solid line represents the predicted likelihood of receiving adjuvant radiation. The dashed lines represent 95% confidence intervals. The inset table represents probability of radiotherapy and 95% confidence interval by tumor size.
In the sensitivity analyses, we explored different numbers of knots in the flexible parametric model to assess the difference in 5-year RMST. The results indicated that changes in the number of knots did not significantly affect the overall findings, which continued to show no significant life expectancy difference. The likelihood ratio test revealed that the 3-knot model was the most well-fitted model. Findings for the 4-knot and 5-knot model are available in Appendix B.
Comment
Principal findings
Our study found that among patients 45 years or younger diagnosed with early-stage cervical cancer, with a tumor measuring 4 cm in diameter or smaller, undergoing fertility-sparing surgery was associated with similar survival outcomes within 5 years of diagnosis compared with standard surgery. While some guidelines 5,6 consider patients with tumors larger than 2 cm to be ineligible for fertility-sparing surgery, we found no evidence that fertility-sparing surgery was associated with inferior life expectancy compared with standard surgery at any tumor size up to 4 cm. However, because fertility-sparing surgery was progressively less common with increasing tumor size, we could not rule out the possibility that fertility-sparing surgery resulted in clinically meaningful differences in life expectancy among individuals with large lesions.
As expected, we also observed that among patients who underwent fertility-sparing surgery, larger tumor size was associated with increasing likelihood of receiving postoperative radiotherapy. Most patients (825 of 897 patients) who received fertility-sparing surgery did not receive adjuvant radiation, even for tumors up to 4 cm.
Results in the context of what is known
This is the first study, to our knowledge, to assess how the association between fertility-sparing versus standard surgery and life expectancy vary as a function of tumor size among patients with early-stage cervical cancer without making strong assumptions about the shape of the survival difference curve or the specific tumor size at which a difference in risk might occur. Thus, this study introduces two significant innovations in the context of the existing literature. First, while prior studies used cut-points in tumor size, we analyzed tumor size as a continuous variable, avoiding size categories that might obscure the underlying relationship between life expectancy and tumor size and reduce precision. Second, we used 5-year life expectancy as our study endpoint, which is a more clinically relevant and understandable measure of survival than more commonly reported hazard ratios.
Bentivegna et al.31 performed a systematic review of oncologic outcomes after fertility-sparing surgery in stage I cervical cancer and proposed a treatment algorithm wherein tumors >2 cm may be treated with neoadjuvant chemotherapy followed by fertility-sparing surgery with an overall recurrence risk of 6–7%. However, the rationale for the 2-cm cutoff was not specifically addressed and instead, Bentivegna et al. relied on historical data from Dargent’s procedure (a vaginal radical trachelectomy with laparoscopic pelvic lymphadenectomy) that resulted in a recurrence risk of 17% in the pooled cohort of for those with tumors >2cm. This systematic review was updated by Morice et al.,32 again demonstrating that for tumors 2–4 cm, the highest recurrence rate (20.7%) was seen among 101 patients who underwent vaginal radical trachelectomy, whereas a recurrence rate of <5% (4.8%) was seen among 375 patients who underwent abdominal radical trachelectomy. Again, however, the study did not evaluate the 2-cm cut point itself and instead separated tumors categorically to <2cm and 2–4cm. Similarly, another study of 733 patients from 44 international institutions found a 3-fold higher recurrence risk after fertility-sparing treatment (19.4% vs 5.7%) among patients with tumors >2 cm compared to those with tumors ≤2 cm, though 25% of the recurrences were cervical only.33
The role of neoadjuvant chemotherapy among patients with larger cervical tumors who desire fertility preservation is an important outstanding area of research. The results of a systematic review and meta-analysis of 23 studies involving 205 patients with early-stage cervical cancer who received neoadjuvant chemotherapy showed that 84.8% of 112 women who desired pregnancy after surgery achieved gestation, with an overall recurrence rate of 12.8% and death rate of 2.8%.34 The ongoing CONTESSA trial, in which patients with stage IB2 cervical cancer are receiving neoadjuvant chemotherapy followed by fertility-sparing surgery, will provide an urgently needed prospective evaluation of this treatment strategy. 35
Clinical implications
Our findings suggest that there may be opportunity for clinicians and patients to discuss the possibility of fertility-sparing surgery even when the tumor size is larger than 2 cm among those patients who are strongly motivated to retain fertility, with thorough counseling regarding the possible remaining risk given this and other studies have been retrospective in nature and limited in number of cases. Additionally, acknowledgement and informed discussion of the increased likelihood of adjuvant radiotherapy in this group is necessary if fertility-sparing surgery is to be performed.
Research implications
Optimal management for individuals with tumors >2 cm but ≤4 cm who desire fertility remains uncertain. Our findings suggest that it may be reasonable to include patients with tumors sized 2 to 4cm in future studies of fertility-sparing surgery.
Additionally, while we found no evidence of harm associated with fertility-sparing surgery for any tumor size <4 cm, fertility-sparing surgery was rarely utilized in patients with tumors between 3 and 4 cm, leading to less precise estimates in this group. Our results could be compatible with a small but clinically meaningful negative effect of fertility-sparing surgery on survival. Larger studies and meta-analyses would be required to estimate the association of fertility-sparing surgery and survival more accurately.
Strengths and limitations
This study was performed using a large and validated dataset representing actual practice patterns across the United States. The fertility-sparing and standard surgery groups were matched on patient, tumor, and treatment-center factors to attempt to minimize differences between the groups. We made no assumptions about the relationship between tumor size and life expectancy outcomes, allowing us to investigate whether the data supported a specific tumor size cutoff for fertility-sparing surgery. There are also important limitations to this study. Despite matching on many measured covariates, the results could be biased because of confounding by unmeasured variables, including tumor-specific factors (lymphovascular space invasion, degree of cervical stromal invasion, and parametrial involvement), patient-specific factors (body mass index, smoking status, HIV status, parity), and surgeon-factors (expertise, decision-making, surgical approach). This study did not include patients who had received neoadjuvant chemotherapy, as this is not a standard of practice in the United States, limiting the international generalizability of our results. Additionally, the NCDB lacks information about cancer recurrence, preventing evaluation of this important outcome, and survival data is limited by years of follow-up available in the dataset. Finally, fertility-sparing surgery in the population of individuals with tumor sizes between 2 cm and 4 cm is a rare occurrence, limiting the precision of our results.
Conclusions
This study suggests that the use of an arbitrary 2-cm threshold in guidelines for risk stratification and decision-making among patients with early-stage cervical cancer may mask opportunities to offer fertility-sparing treatment to select patients with tumors larger than 2 cm.
Supplementary Material
AJOG at a Glance:
A. Why was this study conducted?
This study assessed overall survival differences by tumor size between fertility-sparing surgery and standard surgery for early-stage cervical cancers ≤4 modeling tumor size as a continuous variable.
B. What are the key findings?
Survival was similar among patients who underwent fertility-sparing or standard surgery irrespective of tumor size, though for larger tumors we could not rule out a clinically meaningful difference.
C. What does this study add to what is already known?
Fertility-sparing treatment may be an option for select patients with tumors up to 4 cm..
Acknowledgements:
Editorial support was provided by Sunita Patterson of Editing Services, Research Medical Library at MD Anderson Cancer Center.
Sources of funding:
This work was supported by grants from the National Institutes of Health/National Cancer Institute (NIH/NCI) under Cancer Center Support Grant number P30 CA016672 (J.A. Rauh-Hain, K. Jorgensen, A. Kanbergs, and C-F. Wu); grant number K08 CA234333 (J.A. Rauh-Hain); and grant number T32 CA101642 (K. Jorgensen, A. Kanbergs). Alexander Melamed reports funding from the Department of Defense Ovarian Cancer Research Program (AM OC210024). The funding sources were not involved in the development of the research hypothesis or study design, data analysis, or manuscript writing.
Footnotes
Conflicts of Interest:
A.M. served on an AstraZeneca advisory board. J.A.R.-H. received consulting fees from Schlesinger Group and Guidepoint. P.T.R. received support for travel to the European Society of Gynecological Oncology Congress (2022), International Federation of Gynecology and Obstetrics Regional Meeting (2022), and International Gynecologic Cancer Society Annual Meeting (2022). N.A. received support from Fundación Alfonso Martin Escudero to obtain international research training at MD Anderson Cancer Center. All other authors report no conflict of interest.
CRediT author statement:
Kirsten Jorgensen: Conceptualization, Data curation, Methodology, Writing original draft & editing. Núria Agustí: Conceptualization, Methodology, Writing original draft & editing. Chi-Fang Wu: Data curation, Formal analysis, Methodology, Validation, Writing original draft & editing. Alexa Kanbergs: Writing original draft & editing. Rene Pareja: Methodology, Writing original draft & editing. Pedro Ramirez: Methodology, Writing original draft & editing. Jose Alejandro Rauh-Hain: Conceptualization, Methodology, Supervision, Writing original draft & editing. Alexander Melamed: Conceptualization, Data curation, Methodology, Supervision, Writing original draft & editing.
References
- 1.Shahmoradi Z, Damgacioglu H, Clarke MA, et al. Cervical Cancer Incidence among US Women, 2001–2019. JAMA. 2022;328(22):2267–2269. doi: 10.1001/jama.2022.17806 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Arbyn M, Weiderpass E, Bruni L, et al. Estimates of incidence and mortality of cervical cancer in 2018: a worldwide analysis. Lancet Glob Health. 2020;8(2):e191–e203. doi: 10.1016/S2214-109X(19)30482-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Letourneau JM, Ebbel EE, Katz PP, et al. Pretreatment fertility counseling and fertility preservation improve quality of life in reproductive age women with cancer. Cancer. 2012;118(6):1710–1717. doi: 10.1002/cncr.26459 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Deshpande NA, Braun IM, Meyer FL. Impact of fertility preservation counseling and treatment on psychological outcomes among women with cancer: A systematic review. Cancer. 2015;121(22):3938–3947. doi: 10.1002/cncr.29637 [DOI] [PubMed] [Google Scholar]
- 5.National Comprehensive Cancer Network (NCCN). Cervical cancer (Version 1.2024). Available: https://www.nccn.org/professionals/physician_gls/pdf/cervical.pdf [Accessed 02 January 2024]
- 6.Cibula D, Raspollini MR, Planchamp F, et al. ESGO/ESTRO/ESP Guidelines for the management of patients with cervical cancer - Update 2023. Int J Gynecol Cancer. 2023;33(5):649–666. doi: 10.1136/ijgc-2023-004429 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Wright JD, Matsuo K, Huang Y, et al. Prognostic Performance of the 2018 International Federation of Gynecology and Obstetrics Cervical Cancer Staging Guidelines. Obstet and Gynecol. 2019;134(1):49–57. doi: 10.1097/AOG.0000000000003311 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Ramirez PT, Pareja R, Rendón GJ et al. Management of low-risk early-stage cervical cancer: Should conization, simple trachelectomy, or simple hysterectomy replace radical surgery as the new standard of care? Gynecol Oncol. 2014;132(1):254–259. doi: 10.1016/j.ygyno.2013.09.004 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Schmeler KM, Pareja R, Lopez Blanco A, et al. ConCerv: a prospective trial of conservative surgery for low-risk early-stage cervical cancer. Int J Gynecol Cancer. 2021;31(10):1317–1325. doi: 10.1136/ijgc-2021-002921 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Wright JD, Nathavithrana R, Lewin SN, et al. Fertility-conserving surgery for young women with stage IA1 cervical cancer: safety and access. Obstet Gynecol. 2010;115(3):585–590. doi: 10.1097/AOG.0b013e3181d06b68. [DOI] [PubMed] [Google Scholar]
- 11.Sonoda Y, Abu-Rustum NR, Gemignani ML, et al. A fertility-sparing alternative to radical hysterectomy: How many patients may be eligible? Gynecol Oncol. 2004;95(3):534–538. doi: 10.1016/j.ygyno.2004.07.060 [DOI] [PubMed] [Google Scholar]
- 12.Diaz JP, Sonoda Y, Leitao MM, et al. Oncologic outcome of fertility-sparing radical trachelectomy versus radical hysterectomy for stage IB1 cervical carcinoma. Gynecol Oncol. 2008;111(2):255–260. doi: 10.1016/j.ygyno.2008.07.014 [DOI] [PubMed] [Google Scholar]
- 13.Lintner B, Saso S, Tarnai L, et al. Use of abdominal radical trachelectomy to treat cervical cancer greater than 2 cm in diameter. Int J Gynecol Cancer. 2013;23(6):1065–1070. doi: 10.1097/IGC.0b013e318295fb41 [DOI] [PubMed] [Google Scholar]
- 14.Li X, Li J, Jiang Z, Xia L, Ju X, Chen X, Wu X. Oncological results and recurrent risk factors following abdominal radical trachelectomy: an updated series of 333 patients. BJOG. 2019;126(9):1175. doi: 10.1111/1471-0528.15794 [DOI] [PubMed] [Google Scholar]
- 15.Li J, Wu X, Li X, Ju X. Abdominal radical trachelectomy: Is it safe for IB1 cervical cancer with tumors ≥2 cm. Gynecol Oncol. 2013;131(1):87–92. doi: 10.1016/j.ygyno.2013.07.079 [DOI] [PubMed] [Google Scholar]
- 16.Kato T, Takashima A, Kasamatsu T, et al. Clinical tumor diameter and prognosis of patients with FIGO stage IB1 cervical cancer (JCOG0806-A). Gynecol Oncol. 2015;137(1):34–39. doi: 10.1016/j.ygyno.2015.01.548 [DOI] [PubMed] [Google Scholar]
- 17.Fedewa SA, Cokkinides V, Virgo KS, et al. Association of insurance status and age with cervical cancer stage at diagnosis: National cancer database, 2000–2007. Am J Public Health. 2012;102(9):1782–1790. doi: 10.2105/AJPH.2011.300532 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Practice Committee of the American Society for Reproductive Medicine. Aging and infertility in women. Fertil Steril. 2006. Nov;86(5 Suppl 1):S248–52. doi: 10.1016/j.fertnstert.2006.08.024 [DOI] [PubMed] [Google Scholar]
- 19.Nitecki R, Woodard T, Rauh-Hain JA. Fertility-Sparing Treatment for Early-Stage Cervical, Ovarian, and Endometrial Malignancies. Obstet Gynecol. 2020. Dec;136(6):1157–1169. doi: 10.1097/AOG.0000000000004163. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Edge SB, Byrd DR, Compton CC, et al. Cervix uteri. In: Edge SB, Byrd DR, Compton CC, et al, eds. AJCC Cancer Staging Manual. 7th ed. Springer; 2010:395 [Google Scholar]
- 21.Tong GJ, Zhang GY, Liu J, et al. AJCC Cancer Staging Manual. 8th ed. Springer; 2017:148–161. [Google Scholar]
- 22.World Health Organization. International Classification of Diseases for Oncology (ICD-O). 3rd ed., 1st revision. World Health Organization; 2013. Accessed September 25, 2022. https://apps.who.int/iris/handle/10665/96612 [Google Scholar]
- 23.Royston P, Parmar MKB. Restricted mean survival time: an alternative to the hazard ratio for the design and analysis of randomized trials with a time-to-event outcome. BMC Med Res Methodol. 2013;13:152. doi: 10.1186/1471-484 2288-13-152 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Uno H, Claggett B, Tian L, et al. Moving beyond the hazard ratio in quantifying the between-group difference in survival analysis. Journal of Clinical Oncology. 2014;32(22):2380–2385. doi: 10.1200/JCO.2014.55.2208 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Han K, Jung I. Restricted Mean Survival Time for Survival Analysis: A Quick Guide for Clinical Researchers. Korean J Radiol. 2022;23(5):495–499. doi: 10.3348/kjr.2022.0061 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Jorgensen K, Meernik C, Wu CF, et al. Disparities in Fertility-Sparing Treatment and Use of Assisted Reproductive Technology After a Diagnosis of Cervical, Ovarian, or Endometrial Cancer. Obstet and Gynecol. 2023;141(2):341–353. doi: 10.1097/AOG.0000000000005044 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Nasioudis D; Rubin E; Kim S, et al. Disparities in the use of fertility-sparing surgery for early stage cervical cancer. Fertil Steril. 2018. Sept;110(4 Suppl) E175–176. DOI: 10.1016/j.fertnstert.2018.07.521 [DOI] [Google Scholar]
- 28.Zreik J, Takagi MA, Akhter MF, Ahmad AA, Pandya K, Madoun JM, Bailey B. Sociodemographic Disparities in the Diagnosis and Prognosis of Patients With Cervical Cancer: An Analysis of the Surveillance, Epidemiology, and End Results Program. Cureus. 2023. Jul 6;15(7):e41477. doi: 10.7759/cureus.41477. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Cohen CM, Wentzensen N, Castle PE, et al. Racial and Ethnic Disparities in Cervical Cancer Incidence, Survival, and Mortality by Histologic Subtype. J Clin Oncol. 2022;41:1059–1068. doi: 10.1200/JCO.22 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Hinchliffe SR, Lambert PC. Flexible parametric modelling of cause-specific hazards to estimate cumulative incidence functions. BMC Med Res Methodol. 2013;13:13. doi: 10.1186/1471-2288-13-13 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Bentivegna E, Gouy S, Maulard A, Chargari C, Leary A, Morice P. 508 Oncological outcomes after fertility-sparing surgery for cervical cancer: a 509 systematic review. Lancet Oncol. 2016;17(6):e240–e253. 510 doi: 10.1016/S1470-2045(16)30032-8 [DOI] [PubMed] [Google Scholar]
- 32.Morice P, Maulard A, Scherier S, et al. Oncologic results of fertility sparing surgery of cervical cancer: An updated systematic review. Gynecol Oncol. 2022;165(1):169–183. doi: 10.1016/j.ygyno.2022.01.023 [DOI] [PubMed] [Google Scholar]
- 33.Slama J, Runnebaum IB, Scambia G, et al. Analysis of risk factors for recurrence in cervical cancer patients after fertility-sparing treatment: The FERTIlity Sparing Surgery retrospective multicenter study. Am J Obstet Gynecol. 2023;228(4):443.e1–443.e10. doi: 10.1016/j.ajog.2022.11.1295 [DOI] [PubMed] [Google Scholar]
- 34.Burbano J, Heredia F, Sanabria D, Ardila EG, Viveros-Carreño D, Rodriguez J. Neoadjuvant chemotherapy prior to fertility-sparing surgery in cervical tumors larger than 2 cm: A systematic review on fertility and oncologic outcomes. Int J Gynecol Cancer. 2021;31(3):387–398. doi: 10.1136/ijgc-2020-001765 [DOI] [PubMed] [Google Scholar]
- 35.Plante M, Van Trommel N, Lheureux S, et al. FIGO 2018 stage IB2 (2–4 cm) 528 Cervical cancer treated with Neo-adjuvant chemotherapy followed by fertility 529 Sparing Surgery (CONTESSA); Neo-Adjuvant Chemotherapy and 530 Conservative Surgery in Cervical Cancer to Preserve Fertility (NEOCON-F). 531 A PMHC, DGOG, GCIG/CCRN and multicenter study. Int J Gynecol 532 Cancer. 2019;29(5):969–975. doi: 10.1136/ijgc-2019-000398 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.