We thank Dr Gary S. Collins for his interest and his comments regarding our recent article about the development of a nomogram to predict the 5-year recurrence of borderline ovarian tumors.
The comments suggest that survival-based methods such as Cox regression are the most proper methods to deal with censored observations. We agree totally with this point of view. Indeed, multivariable survival analysis is a well-known tool, as evidenced by the popularity of the Cox model in the medical literature. However, few authors reported that logistic regression may represent an alternative to the analysis of such survival data, when event times are grouped into intervals. Furthermore, it has been shown that results from such an adaptation often will lead to parameter estimates close to those obtained by the proportional hazards model. In addition, several reports have proposed to test factors that are associated with survival in univariate and multivariate analysis according to logistic regression analysis to analyze such survival data.
In the current study, we included all patients who did not experience a recurrence and who were followed for a minimum of 5 years from diagnosis to last follow-up evaluation. Women who experienced a recurrence were included whatever the length of follow up. This naturally means that only women who experienced a recurrence with a follow-up <5 years were also included in the analysis.
We provided a multivariate Cox proportional hazards model on factors that showed significant association on univariate analyses or those that were clinically relevant. International Federation of Gynecology and Obstetrics stage, age at diagnosis, histologic subtype, completeness of surgery, and type of surgery were included in the Cox model ( Table ). The predictive model had a concordance index of 0.72 (95% confidence interval, 0.68–0.76) and 0.68 (95% confidence interval, 0.66–0.70) before and after the 200 repetitions of bootstrap sample corrections, respectively.
| Multivariate analysis | ||
|---|---|---|
| Variables | Hazard ratio (95% confidence interval) | P value |
| Age | 1.14 (0.73–1.78) | .5527 |
| International Federation of Gynecology and Obstetrics staging | .7435 | |
| Ia | Reference | |
| Ib | 0.64 (0.15–2.74) | |
| Ic | 0.70 (0.26–1.91) | |
| II | 0.96 (0.17–5.32) | |
| III | 1.03 (0.47–2.26) | |
| Histologic subtype: serous vs mucinous | 1.11 (0.17–7.25) | .9126 |
| Surgical procedure | .0035 | |
| Bilateral salpingo-oophorectomy ± total hysterectomy | Reference | |
| Unilateral salpingo-oophorectomy | 4.10 (1.59–10.57) | |
| Cystectomy | 11.23 (4.96–25.43) | |
| Completeness of surgical staging: incomplete vs complete | 0.67 (0.27–1.64) | .3781 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
