Introduction

Prostate cancer is the most common cancer for men in North America. In Canada, approximately 24,000 men were diagnosed with prostate cancer and 4100 men deceased from the disease in 2015 [1]. Radical prostatectomy (RP) has been shown to be an effective curative treatment, reducing mortality and metastasis rates [1].

Surgical wait time (SWT) to RP is an important factor to consider. Given budget restraints and ongoing cutbacks, SWT are currently on the rise in Canada. SWT in Ontario had almost doubled between 1980–1995 and 1996–2000 from 55 to 91 days [2]. Prolonged delays have been well documented to increase anxiety; stress and all around decrease their quality of life [3,4,5]. The effect of surgical delay on oncologic prognosis remains a controversial topic and a common question during patient counseling. Association between SWT and biochemical recurrence (BCR) remains equally provocative, with only a limited number of studies reporting a trend towards an increased risk of BCR with surgical delays [6,7,8].

Most analyses performed on such topic were conducted prior to the era of minimally invasive and robotic surgery. There are thus different, novel aspects that are specific to this type of surgery that must be considered. Given that only a few specialized surgeons perform RARP on a limited number of robots in Canada (24 in 2017), concentrated mainly in select metropolitan high-volume centers, and may increase SWT.

Given limitations to resources in a publically funded health care system, we initially explored the impact of surgical wait time (SWT) on adverse pathological outcomes [9]. Herein, we sought to further assess the impact of SWT to robot-assisted radical prostatectomy (RARP) on biochemical recurrence (BCR).

Materials and methods

Patient characteristics

After ethical review board approval, a prospectively collected robot-assisted radical prostatectomy database from two major centers in Montreal (Hôpital du Sacré Cœur de Montréal and Hôpital Saint Luc) was queried to identify all men who underwent RARP between 2006 and 2015. Due to the retrospective nature of the study formal consent is not required. All cases were performed by one of 2 fellowship trained, experienced robotic surgeons using the previously reported technique [10, 11]. Patients who underwent active surveillance (AS) (n = 117), or who were on 5-alpha reductase inhibitors (n = 11) were excluded to accurately assign patients to the correct risk category.

SWT evaluation

Time to surgery was calculated based on the difference between the date of diagnosis by trans-rectal ultrasound (TRUS) prostate biopsy and the date of surgery. Time from biopsy to robotic surgery consultation (date of RARP booking request) and from booking to actual surgery was also calculated for the overall cohort and in each D’Amico risk group. SWT was considered as a continuous variable in the whole analysis.

The primary endpoint in our analysis was biochemical recurrence (BCR), defined as two consecutive PSA ≥ 0.2 ng/dl, or salvage external beam radiation therapy and/or salvage androgen deprivation therapy. BCR-free survival was calculated using the Kaplan–Meier estimator.

The effect of delay on BCR was examined in univariable analysis (UVA) and multivariable analysis (MVA) using the Cox-regression hazard model taking into consideration known prognosticators of BCR. Subgroup analysis was further conducted for all D’Amico risk categories. All statistical tests were 2-sided; R studio software (RSudio Inc, Boston MA, USA) was used in all analyses. A p value < 0.05 was considered statistically significant.

All men were prospectively followed after radical prostatectomy every 4 months for the 1st year, then every 6 months for the 2nd year and on a yearly base thereafter. None of the included patients received adjuvant therapy before proven BCR failure.

Results

Among the identified 687 men eligible for study, after exclusions, 619 had completed demographic, clinical, pathologic, and follow-up data available. Baseline characteristics are presented in Table 1.

Table 1 Baseline characteristics of overall RARP cohort
Full size table

The mean and median follow-up was 28 and 22 months, respectively. Disease-specific (BCR-free) survival was 95.1, 86.7, and 82.4% at 1, 3, and 5 years, respectively.

SWT was significantly different among the 3 D’Amico risk groups with mean SWT of, 169.11 CI (157.02; 181.19), 150.67 CI (143.39; 157.94) and, 125.81 CI (108.82; 142.79) days for low, intermediate and high-risk groups, respectively (p < 0.001). To control for this bias, a subgroup analysis was carried on 3 groups: low, intermediate, and high-risk groups.

Upon analyzing the entire study cohort with MVA, higher pathologic Gleason score (< 0.015), ECE (< 0.013) and positive surgical margin (PSM) (< 0.001), were predictors of BCR. However, SWT did not affect BCR on MVA (p = 0.196) (Table 2).

Table 2 MVA of different predictors of BCR in overall cohort
Full size table

On subgroup analysis, initially reviewing the low-risk group, SWT predicts BCR on UVA p = 0.022 HR = 0.973 (0.950; 0.996). There was no association between SWT and BCR on MVA (p = 0.086) (Table 3).

Table 3 MVA of different predictors of BCR in D’Amico low-risk group
Full size table

With regards to the intermediate risk group, prolonged SWT was not predictive of BCR (p = 0.991) (Table 4). Independent predictors of BCR in the intermediate-risk group were PSM (p = 0.001), and ECE (p = 0.016).

Table 4 MVA of different predictors of BCR in D’Amico intermediate risk group
Full size table

When analyzing the high-risk group SWT significantly predicts BCR on UVA (p = 0.053). SWT (p = 0.001), and PSM (p = 0.027) increase the risk of BCR (p = 0.001) on MVA (Table 5). Of note 71 patients had high-risk cancer.

Table 5 MVA of different predictors of BCR in D’Amico high risk group
Full size table

Receiver operator characteristics (ROC) cut-off analysis showed a value of 90 days. This value was validated on UVA with Cox-regression model, and on MVA.

Indeed, Kaplan–Meier estimation curve was analyzed for high-risk patients above and below 90 days wait time. Comparison of both curves with likelihood test showed significant difference p = 0.03 (Fig. 1).

Fig. 1
figure 1

Kaplan–Meier plot of BCR-free survival for the high-risk group, stratified according to SWT 90 days cut-off

Full size image

Discussion

The present study suggests the fact that SWT are prolonged in 2 Canadian academic centers. This fact is comparable to other reported Canadian centers, as already demonstrated in previous reports [2, 9, 12,13,14,15]. Indeed, in 2006, a Canadian consortium of experts recommended a maximum wait time of 90, 60 and 28 days in low, intermediate and high-risk prostate cancer, respectively. The present study showed that wait times largely exceed recommended consensus wait times. On the other hand, it states that it is safe to wait in low- and intermediate-risk groups far above the suggested time frame for low and intermediate risks [16].

Unique to our data is a large, contemporary patient cohort in the era of robotic surgery where a perspective on delays towards surgery in a public health care system, can be further studied. In contrast to our initial publication addressing SWT and its impact on adverse surgical pathology [9], the current study demonstrates that SWT was positively correlated to BCR for high-risk group (p = 0.001), when it exceeds 90 days.

Similar to previously reported studies on surgical delay, SWT was defined as time interval elapsing between biopsy date and surgical intervention [7, 17,18,19,20]. Furthermore, the overall SWT was divided into time from biopsy to booking (overall mean of 79.21 days) and time from booking to surgery (overall mean of 76.12 days). As it was described in an earlier study [9, 12], these time intervals are relatively long reflecting delays in multiple steps of patient management. As already discussed in our previous report on SWT and CAPRA-S pathology [9], our data demonstrate that time between biopsy and booking consult are similar in all three risk group stratification since this period is not dependent on the surgeon’s influence. However, there is a significant difference in delay from booking consult to treatment, suggesting the influence of surgeons to operate patients with higher risk faster.

In the present study, SWT did not significantly predict BCR in the overall cohort, as was the case in previous report [18, 19, 21,22,23,24]. This holds also true, for low- and intermediate-risk group, once we divided the cohort with the D’Amico risk stratification, which is also concordant with most of the previous studies [7, 22, 25,26,27].

On the other hand, two studies based on D’Amico low-risk patients found that delays greater than 6 months were associated with a significantly higher risk for BCR [8, 28].

Indeed, O’Brien demonstrated that delays above 6 months were associated with higher upgrading (47 vs 27% upstaging to Gleason score 7–10) and higher BCR rates [8]. Same cut-off (6 months) was also reported by Freedland in low-risk category and showed higher BCR rates [28]. Two other studies incorporated also intermediate risk patients and found positive association. Abern et al. analyzed intermediate risk men (D’Amico stratification) and noted that delay above 9 months was associated with higher rates of PSM and BCR [7]. A threshold of 19.2 months was defined by Holmstrom et al. to be associated with higher Gleason grade compared to immediate surgery (3.5 months). No association with BCR was found [29].

Our study strengthens the concept that it is safe to wait for low- and intermediate-risk groups, matching previous reports that found negative association of SWT with BCR.

Previous studies failed to show positive association of SWT to BCR, for high-risk patient operated of radical prostatectomy. The only study validating association of delayed wait time to BCR in high risk patient (T ≥ 2b, PSA > 10, Gleason > 6, > 34–50% positive biopsy cores) was that done by Nguyen but studied exclusively radiation therapy patients [6]. Indeed, delays above 2.5 months were found to be independent predictor of time to BCR. Three series incorporated high-risk patients and found no association [25, 26]. On the other hand, Nam et al. included high-risk patient and found positive association for overall, but also disappeared after adjustment. They concluded that negative results could be due to retrospective bias and limited number of patient [27]. The present study showed that in the high-risk group, waiting beyond 90 days to perform RARP is associated with higher BCR. This holds true after MVA. Based on our results, and in the lack of prospective reports, or retrospective reports with larger number of cases, we recommend operating on patient with high-risk disease before 90 days.

Finally, we need to shine the light on the fact that the present study as well as most of the previous ones, evaluated pathological outcomes or BCR as surrogates for CSS and OS. Only three studies evaluated the effect of wait time on CSS and OS, where no association was found [20, 29, 30].

Our study gives a unique perception on the limitations of universal health system like Canada may have on SWT. This is the first study to show positive association of SWT with BCR for patient operated exclusively by the robot. It defines a cut-off of 90 days.

Study limitations include its retrospective nature, which can include inherent selection biases. From a statistical point of view, inside the high-risk group, number of individuals was relatively limited compared to the number of variables considered. Furthermore, number of high-risk patients was relatively low compared to low- and intermediate-risk patients. We also acknowledge the fact that wait-time is partly dependent on the urologists’ preference and tried to compensate on this selection bias by stratifying the patients by the D’Amico risk groups and adjusting in multivariable analysis. Follow-up time was relatively short concerning the development of BCR especially in low- and intermediate-risk groups. It is also subject to time to event bias depending on the definition of wait time as time from biopsy to BCR.

Conclusion

In the present study, we evaluated SWT for Canadian men in a publically funded, universal healthcare system, its variation between D’Amico risk categories and impact on BCR. Based on our findings, it appears that SWT should be less than 90 days for men with high-risk disease. While surgeon case-selection appears to influence SWT, other factors also require closer evaluation to improve timing to definitive prostate cancer treatment. Further studies are warranted to validated the suggested cut-off and to assess the impact of SWT on cancer-specific survival and overall survival.