Impact of Median Lobe Anatomy: Does Its Presence Affect Surgical Margin Rates During Robot-Assisted Laparoscopic Prostatectomy?

Introduction

T he presence of a median lobe during robot-assisted laparoscopic prostatectomy (RALP) for prostate cancer can pose technical difficulties during dissection because of distortion of bladder neck anatomy with ergonomic and visual obstruction during antegrade dissection of the prostate. Sarle and associates ¹ reported these difficulties early in their experience when their group was pioneering RALP techniques.

Since then, only two groups have published data suggesting equivalent oncologic outcomes during RALP, despite these difficulties. The main outcome in these studies was positive surgical margin rates (PSMR). In the first single-surgeon cohort of 58 patients, of whom 29 had median lobes, Jenkins and colleagues ² found no difference in PSMR. In another single-surgeon cohort of 154 patients, of whom 29 had median lobes, Meeks and coworkers ³ also found no difference in PSMR.

Unfortunately, the current body of evidence is arguably not robust. Therefore, it is difficult to determine the exact relationship of median lobe anatomy and PSMR in this condition that is not infrequently encountered by the robotic urologic surgeon. Neither of the previous studies included multivariable analysis. Gleason score, pathologic stage, and prostate weight are known predictors of positive surgical margins during RALP and should be considered when analyzing the effect of median lobe anatomy. ^{4 –6} Also, both studies analyzed small cohorts involving a single surgeon at an academic institution. It has been suggested that surgeon experience and training may affect PSMR. ^7,8

The purpose of this study was to expand on previous studies by testing the hypothesis that median lobe anatomy does not affect PSMR in a larger cohort of patients involving a multisurgeon robotics program. Furthermore, we measured the impact of median lobe anatomy through multivariable analysis.

Patients and Methods

This study was approved by our Institutional Review Board. A total of 1022 patients underwent RALP at our institution between August 2008 and December 2010 by 12 surgeons. This cohort represents our institution's initial experience with RALP, as described previously. ⁹ Of the 1022 subjects, 791 had known median lobe anatomy that was prospectively recorded at the time of the operation. The presence of a median lobe (yes or no) was judged and reported by the operating surgeon. Age, race, body mass index, American Society of Anesthesiologists score, PSA level, Gleason score, pathologic stage, prostate weight, surgical margin status, estimated blood loss (EBL), total operating room time, and intraoperative and postoperative complications were collected prospectively.

Results

Table 1 summarizes the demographic and clinical characteristics of the cohort overall and by median lobe status. Compared with patients without a median lobe (n=672, 85.0%), patients with a median lobe (n=119, 15.0%) were less likely to have positive surgical margins (16.0% vs 24.4%; P=0.044). They were older (median 63 y vs 60 y; P=0.001) and had higher PSA levels (median, 6.1 ng/dL vs 5.4 ng/dL; P=0.003). They were also more likely to have lower Gleason scores (6, 58.1% vs 42.1% ; P=0.012), less aggressive pathologic stage (pT₂, 87.4% vs. 75.4%; p=0.024), and larger prostates (median weight, 64 g vs 48 g; P<0.0001). They experienced higher EBL (>300 mL, 8.4% vs 4.2%; P=0.004), but total operative time and rate of intraoperative and postoperative complications were not significantly different between the two groups.

Analysis of PSMR was stratified by surgical site of positive margins. At each of the five surgical sites of the prostate—apex, base, anterior, lateral, and posterior—prevalence of positive margins was compared between patients with and without a median lobe (Table 2). The median lobe patients were less likely to have positive margins posteriorly (21% vs 47%; P=0.034).

The RR for positive surgical margins that was associated with different factors was estimated with univariable and multivariable Poisson regression models (Table 3). After adjusting for age, prostate weight, PSA, Gleason score, and pathologic stage in the multivariable model, median lobe anatomy did not differ with respect to risk for positive margins (P=0.885). Other risk factors were related to positive margin risk in different directions. Prostate weight was inversely related to the risk, while PSA, Gleason score, and pathologic stage were positively related to the risk.

Discussion

Previous studies suggested no impact of median lobe anatomy on PSMR. ^2,3 Using a significantly larger cohort of 791 patients and a larger number of 119 patients with median lobes, however, we were able to detect a lower PSMR in patients with median lobes (16.0% vs 24.4%; P=0.044). In addition, we demonstrated that PSMR was significantly decreased at the posterior margin in patients with median lobes when compared with other surgical sites (P=0.034).

We theorize several explanations for this finding. The median lobe may force the surgeon to retract the posterior aspect of the prostate more anteriorly during dissection, thereby allowing more exposure to the area. The decreased PSMR posteriorly could confirm the beneficial effects of the technique described by Patel and associates. ¹⁰

By comparing characteristics of patients with and without median lobes in Table 1, we found that patients with median lobes were also older, had higher PSA levels, lower Gleason scores, lower pathologic stages, and larger prostates (P<0.024). In previous reports, lower Gleason score, lower pathologic stage, and larger prostate size predicted lower PSMR. ^{4 –6}

Our multivariable model adjusted for all these characteristics. Although presence of a median lobe suggested a decreased RR for positive margins (RR=0.65, 95% confidence interval 0.42–1.01; P=0.055) in univariable analysis, it was not an independent predictor of PSMR in multivariable analysis. It is likely that the effect of median lobe anatomy on lower PSMR is actually a result of its association with lower Gleason scores, lower pathologic stages, and larger prostates. Our multivariable analysis of these specific characteristics echoed previous reports with the same findings of decreased RR with each variable. ^{4 –6}

Despite the fact that our results did not show the presence of the median lobe being an independent predictor of PSMR, we did show in our large cohort that the median lobe is associated with favorable characteristics that may decrease PSMR. This may allow the robotic surgeon to more accurately predict outcomes of a patient when a median lobe is encountered.

Moreover, our results may reflect more accurately that of a community urology robotics program because our results are collected from a diverse group of 12 surgeons with varied experience and training. This is in contrast to previous studies, where single high-volume surgeons were examined. ^2,3 It is known that surgeon experience and training can affect PSMR. ^7,8

There are several limitations to this study. First, 22.6% (231/1022) of subjects in the consecutive cohort did not have known median lobe anatomy. There was a 15% prevalence of median lobes in our cohort of subjects with known anatomy, however, which is similar to 18%, the prevalence of median lobes found in a previous consecutive cohort. ³ The similar rate suggests that the missing data can be assumed to be from random errors in prospective collection.

Second, the presence of a median lobe was based on subjective surgeon perspective. Furthermore, presence was a binary definition, and actual size and character of the median lobe was not addressed. It is our assumption, however, that in our large median lobe cohort, the extremes in size and character of the median lobes are minimized in our results.

Finally, time to urinary continence was not assessed in this study. The bladder neck defect resulting from median lobe dissection can be larger, necessitating reconstruction and threatening postoperative continence. Of note, Jenkins and coworkers ² and Meeks and colleagues ³ did not detect any differences in postoperative continence rates between patients with and without median lobes. A future study on this cohort analyzing bladder neck reconstruction and continence rates can contribute to a more complete assessment of the impact of median lobe anatomy on RALP outcomes.

Conclusion

Although median lobe anatomy is not an independent predictor of positive surgical margins in RALP, it is associated with favorable pathologic characteristics that are known to predict decreased positive surgical margins.