Role of preoperative magnetic resonance imaging on the surgical outcomes of radical prostatectomy: Does preoperative tumor recognition reduce the positive surgical margin in a specific location? Experience from a Thailand prostate cancer specialized center

Thitipt Hnsomwong ＊, Pt Sksirismpnt ,Sudhir Ishrwl , Puordee Aussvvirojekul Vrt Wornisrkul Siros Jitprphi Suni Leewnsngtong Twthi Tweemonkongsp Sittiporn Srinulnd

a Division of Urology,Department of Surgery,Faculty of Medicine Siriraj Hospital,Mahidol University,Bangkok, Thailand

b Division of Urology, Department of Surgery, Somdech Phra Pinklao Hospital, Naval Medical Department, Royal Thai Navy, Bangkok, Thailand

c Department of Urology, Oregon Health and Science University, Portland, OR, United States

KEYWORDS Preoperative magnetic resonance imaging;Prostate cancer;Positive surgical margin;Radical prostatectomy;Apex;Apical positive surgical margin

Abstract Objective: Multiparametric magnetic resonance imaging (MRI) has become the standard of care for the diagnosis of prostate cancer patients.This study aimed to evaluate the influence of preoperative MRI on the positive surgical margin (PSM) rates.Methods: We retrospectively reviewed 1070 prostate cancer patients treated with radical prostatectomy (RP) at Siriraj Hospital between January 2013 and September 2019.PSM rates were compared between those with and without preoperative MRI.PSM locations were analyzed.Results: In total,322(30.1%)patients underwent MRI before RP.PSM most frequently occurred at the apex (33.2%), followed by posterior (13.5%), bladder neck (12.7%), anterior (10.7%),posterolateral (9.9%), and lateral (2.3%) positions.In preoperative MRI, PSM was significantly lowered at the posterior surface (9.0% vs.15.4%, p=0.01) and in the subgroup of urologists with less than 100 RP experiences(32%vs.51%,odds ratio=0.51,p<0.05).Blood loss was also significantly decreased when a preoperative image was obtained(200 mL vs.250 mL,p=0.02).Multivariate analysis revealed that only preoperative MRI status was associated with overall PSM and PSM at the prostatic apex.Neither the surgical approach, the neurovascular bundle sparing technique, nor the perioperative blood loss was associated with PSM.Conclusion: MRI is associated with less overall PSM, PSM at apex, and blood loss during RP.Additionally, preoperative MRI has shown promise in lowering the PSM rate among urologists who are in the early stages of performing RP.

1.Introduction

Multiparametric magnetic resonance imaging (mpMRI) is integrated into clinical pathways to diagnose patients with prostate cancer (PCa) after elevated screening prostate-specific antigen (PSA) levels.Prior studies have reported an improved detection rate of clinically significant PCa with mpMRI[1,2].Urologists and radiologists utilize mpMRI to classify lesions into five categories according to Prostate Imaging Reporting and Data System version 2.1(PI-RADS v2.1), of which the categories 4 and 5 are highly suspicious for clinically significant cancer and should be biopsied, while category 3 remains controversial [3].The mpMRI-directed biopsies in selected patients have reduced the over-detection and thus overtreatment of indolent PCa[4-6].

Radical prostatectomy (RP) is considered the standard of care for the treatment of localized PCa [7].RP offers good oncological outcomes in appropriately selected patients and avoidslocalsideeffectsfromionizingradiationandthesystemic side effects of adjuvant androgen deprivation therapy [8].However,RP can lead to long-term side effects such as urinary leakage or sexual dysfunction.The neurovascular bundle(NVB)and urethral length are preserved during the surgery to decrease the long-term side effects.However, aggressive nerve-sparing may sometimes lead to positive surgical margin(PSM)and poor long-term cancer control.PSM predicts a poor clinical outcome after RP[9-11].Prior studies have reported the rates of PSMs after RP around 37.3%-45.8%depending on the NVB sparing technique utilized[12].

This study investigated whether the preoperative information of cancer location and prostate anatomy from mpMRI leads to a better surgical margin status during prostate excision.The anatomical details of prostate and PCa with mpMRI preoperatively may enable an individualized surgical approach to tailor the dissection and nerve-sparing while still offering satisfactory cancer control.

2.Patients and methods

The study protocol was approved by the Ethics Committee of Siriraj Institutional Review Board (419/2562 [EC2]),Faculty of Medicine Siriraj Hospital, Mahidol University,Bangkok, Thailand.

We retrospectively reviewed the patients who had undergone RP in our hospital between January 2013 and September 2019 in the study.The patients missing the parameters of interest were excluded.Patients were classified into two groups based on preoperative magnetic resonance imaging (MRI) status.The patients’ characteristics including age (year), body mass index (BMI, kg/m2),initial PSA(iPSA)level(ng/mL),and prostate size(mL)were recorded.Details of the surgery, including the operative time (min), blood loss (mL), surgical approach (open RP[ORP], laparoscopic RP [LRP], and robotic-assisted RP[RARP]), and NVB techniques utilized (none, unilateral sparing, and bilateral sparing), were recorded.The tumor characteristics and PSM locations were evaluated by reviewing the official pathological reports.A high tumor percentage involvement was defined as a tumor amount over 50%of a total prostate specimen.The sites of the PSM were classified as: apex, anterior, bladder neck, lateral,posterior, and posterolateral.Gleason grade group was determined using 2014 International Society of Urological Pathology(ISUP)grading.Pathological tumor stage(pT)was based on the 2016 tumor-node-metastasis staging definition.Surgeons who had performed over 100 RP procedures were considered high-experience surgeons.

Preoperative MRIs were performed either in or outside our hospital.The MRI results and the tumor locations were officially reported following PI-RADS v2.1,which combines anatomic T2W imaging with functional and physiologic assessment, including diffusion-weighted imaging and its derivative apparent diffusion coefficient maps, and dynamic contrast-enhanced MRI.Preoperative prostate biopsy was performed by either MRI transrectal ultrasound(TRUS) fusion biopsy or systemic TRUS biopsy in or outside our hospital.Excision techniques were utilized according to the operating urologists’preference after the MRI result was discussed with the patient.RP was performed by attending staffs,residents,or fellows under the attending supervision.

Demographic data and intraoperative data were reported with the mean and standard deviation.Blood loss,BMI, tumor percentage involvement, and iPSA level were reported with median with interquartile range.Kruskal-Wallis test was used for between-group comparisons.Categorical data were compared using the Chi-square test.Propensity score matching between patients with and without preoperative MRI had been undertaken with known independent predictors of PSM considered (age, BMI, iPSA, prostate volume, tumor percentage involvement, pT, NVB sparing technique, operative time, and perioperative blood loss) before hypothesis testing was commenced.Univariate logistic regression was conducted for all the variables.We used a multivariate logistic regression model utilizing a backward selection method to discover the independent factors associated with PSM.Multivariable analysis was also performed on each location of PSM.

Statistical analyses were performed using the Python statistical package (Python Software Foundation, Wilmington, DE, USA) elements as follows: SciPy version 1.5.2,lifelines version 0.25.10, and Matplotlib version 3.3.2.A p-value of less than 0.05 was considered statistically significant.

3.Results

In total, 1070 men were included in the study.Among these, 322 (30.1%) patients underwent MRI prior to RP.Out of the MRI group, 172 (53.4%) had the investigation within our facility while the rest were from external sources.The patient and tumor characteristics were similar in both groups (Table 1).The mean operative times were not different (182.51 min in preoperative MRI group vs.187.64 min in non-MRI group, p=0.36).The median blood loss was significantly lower in preoperative MRI group(200 mL vs.250 mL,p=0.02).The median iPSA levels were similar in both groups (10.1 ng/mL in preoperative MRI group vs.10.0 ng/mL in non-MRI group,p=0.41).The mean values of prostate size and tumor involvement percentages were comparable (44.07 mL in preoperative MRI group vs.42.77 mL in non-MRI group, p=0.40; 12% in preoperative MRI group vs.15% in non-MRI group, p=0.10).Most of the tumors were pT2 (56.5% in preoperative MRI group and 48.8% in non-MRI group).The surgical approach was different between the two groups: in preoperative MRI group, RARP was the most selected procedure (72.7%),followed by LRP(18.9%)and ORP(8.4%);whereas RARP was performed in 56.0%, followed by 30.9% LRP and 13.1% ORP in the non-MRI group (p<0.05).Nevertheless, the NVB sparing technique and surgeon experience were well balanced in both groups (Table 1).Subgrouping of low experience surgeons found that only 27 cases(37.0%)in the preoperative MRI group were pT3/4, but there were 99 (55.6%) in the non-MRI group.On the other hand, the proportions of pT3/4 were not different in high experience surgeons (43.4% [108/249] in preoperative MRI group vs.48.2% [275/570] in non-MRI group).

Table 1 Clinical and histological characteristics.

PSM occurred most frequently at the apex in both groups (29.8% in preoperative MRI group and 34.6% in non-MRI group, p=0.30), and the posterior (15.4%), bladder neck (13.6%), anterior (10.6%), posterolateral (9.9%), and lateral (2.5%) locations in non-MRI group.Interestingly, in the preoperative MRI group,when the urologist had details of the tumor location,the PSM was significantly decreased at the posterior surface to 9.0% (p=0.01), while the PSM rates were not different at the other locations.Overall rates of PSM were similar in both groups.However,rate of multiple foci PSM was significantly lower in preoperative MRI(18.6%vs.26.9%,p<0.05)(Table 2).Further analysis in patients with PSM showed that PSM rate was reduced from 51% to 32% (p<0.05) in low experience surgeons with preoperative MRI while surgical approach, nerve-sparing technique,or pT stage had no effect on the PSM(Table 3).Although MRI and histopathological location were 85%correlated, 28% of tumors responsible for PSM were not detected.

The propensity score matching selected 275 patients in MRI group and 338 patients in non-MRI group with similar other significant figures.Univariate analysis of the PSM in propensity-matched population showed that the high iPSA,ISUP group, pT, and tumor involvement percentage were associated with an increased PSM (odd ratio [OR]=2.9 [p<0.05] in patients with a iPSA equal to or above 20 ng/mL,OR=8.3[p<0.05]in ISUP 5,OR=2.8[p<0.05]in pT3a, and OR=5.4 [p<0.05] with a high involvement percentage), while preoperative MRI and surgeon experience correlated with the reduced PSM(OR=0.62[p<0.05]in MRI group and OR=0.57 [p<0.05] in low experience surgeon).In contrast to none nerve-sparing technique, the decision on NVB preservation (either unilateral or bilateral)and increased perioperative blood loss (<600 mL,600-1200 mL, or ≥1200 mL) were not individually associated with the PSM.Operative time initially associated with decreased PSM (OR=0.53 [p<0.05] at 180-240 min), but the multivariate analysis showed that it was confounded by other parameters (OR=0.60 [p=0.10]) (Table 4).

Table 2 PSM location.

Subgroup analysis of the PSM at the apex of the prostate revealed that preoperative MRI and longer operative time(180-240 min compared to <120 min) significantly decreased the rate of PSM in these most common surgically positive areas (OR=0.69, p<0.05 and OR=0.51, p<0.05);while aggressive tumor characteristics were still associated with an increased PSM(OR=2.6[p<0.05]in patients with a iPSA ≥20 ng/mL, OR=4.4 [p<0.05] in ISUP 5) (Table 5).However, pT was firstly revealed as a risk factor for PSM at apical area in the univariate analysis (OR=1.80, p<0.05),and did not significantly influence apical PSM in the multivariate analysis when including other factors (OR=0.81,p=0.50).In contrary,subgroup of the posterior PSM showed that MRI tends to reduce the margin in the propensity-matched univariate analysis (OR=0.58,p=0.04), but the multivariate analysis of this parameter was marginally insignificant (OR=0.60, p=0.06).

Table 3 Influences of RP approach and pT on PSM.

Table 4 Univariate and multivariate analyses of the risk of overall PSM in the propensity-matched group.

4.Discussion

All urologists performing RP strive for the best cancer control,no urine leakage,and undisturbed sexual function.In our practice, we have often been concerned with PSM which may lead to early biochemical failure and risk of cancer recurrence.In the past, the best information the surgeons could get preoperatively on the tumor location or size came from a digital rectal exam, TRUS, or biopsy results.

A prior report detailed the risk factors of PSM after RP in a Thai population that a physician could use for patient counseling and to aid the treatment decision[13].Since the emergence of detailed 3 T MRI, insightful information on the tumor configuration,location,and extension related to the nearby pelvic organ structures now allows urologists to do some anatomical information prior to stepping into the operating room.The mpMRI has good sensitivity for the detection and localization of clinically significant PCa,especially when the lesion size is larger than 10 mm [14].This preoperative information could help select the NVB sparing technique,apical dissection plan,and bladder neck preservation strategy.The current literature reveals that preoperative MRI may alter urologists’ choices toward a more aggressive surgical excision strategy in 29%-39% of cases [15,16].

Table 5 Univariate and multivariate analyses of the risk of apical PSM in the propensity-matched group.

In our center, the apex was the most common site of PSM.This aligned with the previous studies stating that up to 65.4% of patients had adenocarcinoma located at the apex [17,18].Aggressive efforts to preserve the urethral length may cause a PSM at the apex since maximization of this length is associated with improved urinary continence[19].Migration in the prostate biopsy technique may also be responsible for the PSM at the apical area.There was a change in prostate biopsy methods in our center with the introduction of the MRI TRUS fusion biopsy technique.The majority of prostate biopsies in recent years have been performed transperineally with fusion biopsy, while we have routinely performed this transrectally in the past decade.As a result, we found that transperineal biopsy leads to difficulty when dissecting the prostate apex during RP because of the unpreceded report adhesion [20,21].Lastly, the prostate capsule is poorly defined at the prostate apex and possibly leads to increased number of PSM reports in this area [22].

For the posterior PSM, even though our study showed insignificant differences in PSM rate of the location, few studies reported the beneficial effect of preoperative MRI on posterior surface margin status due to visibility and degree of freedom in surgical resection.Wibulpolprasert et al.[23] showed that MRI had the highest sensitivity of 83.1% for detection of index PCa lesions in the midgland area reflecting on visibility on the posterior of the prostate,whereas sensitivity at the apex and the bladder were 71.4%and 64.0%, respectively.Ja¨derling et al.[24] showed that preoperative MRI mostly reduced PSM at the midportion surface since the thick Denonvilliers’ fascia and perirectal fat at the posterior surface may allow surgeons to come up with varied surgical plans to achieve less PSM.

Our report reiterated that tumor aggressiveness is an important predictor of PSM.The pT3,a poorly differentiated tumor (ISUP≥2), high tumor involvement percentage(≥50%), and high PSA (≥10 ng/mL) were all associated with the overall PSM of the prostate.According to previous studies on preoperative MRI, both 1.5 T and 3 T did not significantly improve this oncological outcome [15,25].Rud et al.[15] suggested that more extended excision may be required to reduce PSM even in the setting of acquired preoperative MRI.Unlike previous reports,preoperative MRI led to a significant lower PSM particularly in low volume surgeons with RP experience of under 100 cases in this study.Blood loss was also reduced in the patients who obtained preoperative MRI.This observation might result from the perception of prostate contour leading to the superior prostate dissection plan.Previous studies showed that prostate size and prominent apical periprostatic venous complex revealed by the MRI were associated with increased blood loss during the RP[10,26].Moreover,the result of the present study demonstrated that apical PSM was also significantly reduced when MRI was performed before the prostatectomy, regardless of the surgeon’s experience in performing RP.

The NVB sparing technique is an important part of RP.Meticulous dissection into the correct plane between the prostate capsule and levator ani fascia for cavernous nerve preservation leads to a superior functional outcome and quality of life, namely related to continence and erectile function [27,28].One of the major concerns about using preoperative MRI to direct the RP strategy is that the imaging can misguide the surgeon to either over- or under-resection of the NVB.According to a recent meta-analysis, adding preoperative MRI into the surgical template decision process is appropriate in 77% of cases.In other words, 23% of decisions based on MRI are wrong and lead to PSM or unnecessary NVB removal [29].The authors found that the NVB sparing technique was not associated with higher PSM.Both unilateral and bilateral approaches did not affect the PSM rate at the posterolateral surface or even at the apex,compared to wider non-cavernous nerve-sparing resection.Nevertheless, we did not include the data on preoperative MRI-misguided unnecessary resections of NVB in this study.

Our study is not without limitations.It was a retrospective study that is inherently prone to selection bias,although the baseline groups were similar in characteristics and propensity score matching was undertaken prior to hypothesis testing.Furthermore, there were differences in the preoperative MRI protocol in our study.MRI performed outside of our center may include 1.5 T mpMRI without rectal coil, which may lead to an inadequate detection of tumors and may have affected the margin status of the cases we studied.

5.Conclusion

MRI is currently a valuable investigation for the diagnosis and treatment of PCa.This imaging system helps to decrease overall PSM, PSM at apex, and blood loss.Additionally, preoperative MRI has shown promise in lowering the PSM rate among urologists who are in the early stages of performing RP.

Author contributions

Study concept and design: Thitipat Hansomwong, Varat Woranisarakul.

Data acquisition: Pat Saksirisampant, Siros Jitpraphai,Sunai Leewansangtong.

Data analysis: Pubordee Aussavavirojekul.

Drafting of manuscript: Thitipat Hansomwong, Pubordee Aussavavirojekul.

Critical revision of the manuscript: Sudhir Isharwal,Tawatchai Taweemonkongsap, Sittiporn Srinualnad.

Conflicts of interest

The authors declare no conflict of interest.