Fusion Biopsy Improves Detection of Prostate Cancer
Multiparametric Magnetic Resonance Imaging-Ultrasound Fusion Biopsy Improves but Does Not Replace Standard Template Biopsy for the Detection of Prostate Cancer
The Journal of Urology
Nawar Hanna, Matthew F. Wszolek, Amirkasra Mojtahed, Edouard Nicaise, Bo Wu, Francisco J. Gelpi-Hammerschmidt, Keyan Salari, Douglas M. Dahl, Michael L. Blute, Mukesh Harisinghani and Adam S. Feldman
Purpose: There exists a growing debate as to whether multiparametric magnetic resonance imaging with fusion transrectal ultrasound guided prostate biopsy alone without a standard template biopsy is sufficient to evaluate patients with suspected prostate cancer. Our objective was to describe our experience with fusion targeted prostate biopsy and assess whether it could obviate the need for concomitant standard 12-core template prostate biopsy.
Materials and Methods: We retrospectively reviewed our prospectively collected database of patients who underwent fusion transrectal ultrasound guided prostate biopsy. All images and lesions were graded according to the Prostate Imaging Reporting and Data System, version 2. All patients underwent targeted biopsy followed by standard 12-core double sextant biopsy within the same session. Clinically significant prostate cancer was defined as Grade Group 2 or greater prostate cancer.
Results: A total of 506 patients were included in analysis. Indications were elevated prostate specific antigen with a previous negative prostate biopsy in 46% of cases, prostate cancer on active surveillance in 35%, elevated prostate specific antigen without a prior prostate biopsy in 15% and an isolated abnormal digital rectal examination in 3%. For standard vs fusion prostate biopsy the overall cancer detection rate was 57.7% vs 54.0% (p=0.12) and the clinically significant prostate cancer detection rate was 24.7% vs 30.8% (p=0.001). Of the 185 patients diagnosed with clinically significant prostate cancer 29 (16%) would have been missed if only targeted fusion prostate biopsy had been performed.
Conclusions: Fusion targeted prostate biopsy is associated with a higher detection rate of clinically significant prostate cancer compared to standard double sextant biopsy. However, standard double sextant biopsy should still be performed as part of the routine fusion targeted prostate biopsy procedure to avoid missing a significant proportion of clinically significant prostate cancer.
Prostate cancer is a common disease in the United States with estimated 180,000 new cases in 2016. 1The current gold standard diagnostic test consists of a 12-core double sextant TRUS guided PBx of the peripheral zone. This procedure has been associated with a nonnegligible false-negative rate and under grading as well as over diagnosis of low grade, clinically insignificant PCa. 2,3
In the last decade mpMRI has been increasingly used for the focal identification of PCa and it was suggested that it reliably identifies csPCa. 4,5The exact role and indication of mpMRI in the management of suspected PCa or PCa remains to be clearly defined. 6,7There exists growing debate as to whether mpMRI targeted biopsy alone without standard template biopsy is sufficient to evaluate patients with target lesions on mpMRI. 6,7High variability has been reported in the rate of csPCa found on systematic biopsies and missed by MRI targeted biopsy. 8,9
We report our outcomes of targeted fusion biopsy followed by concomitant double sextant biopsy. We investigated the relative usefulness of targeted fusion biopsy and standard sextant biopsy to detect PCa.
Materials and Methods
With institutional review board approval we retrospectively reviewed a prospectively collected database of all patients who underwent TRUS guided fusion PBx at our institution from July 2014 to January 2018. In patients eligible for fusion biopsy 1 or more lesions were identified on prior prostate mpMRI. Indications for prostate mpMRI were at the discretion of the clinicians. They were categorized as 1) elevated prostate specific antigen with a prior negative biopsy, 2) PCa during active surveillance, 3) elevated PSA without a prior biopsy or 4) abnormal digital rectal examination with normal PSA.
Magnetic Resonance Imaging Acquisition and Interpretation
MRI studies were performed at Massachusetts General Hospital at least 6 weeks after any previous biopsy, if applicable. Studies were performed using a 1.5 Tesla Signa HDxt magnet or a 3.0 Tesla Discovery MR750 magnet (GE Healthcare, Waukesha, Wisconsin) with or without an endorectal coil. In February 2015 our institution made a programmatic shift to the 3.0 Tesla magnet without an endorectal coil.
The MRI protocol included triplane T2-weighted, axial diffusion-weighted and axial dynamic contrast-enhanced sequences. The prostate gland was segmented using DynaCAD software (Invivo™) to determine gland volume. Regions of interest were also drawn around identified lesions for targeting on subsequent UroNav® fusion PBx.
All patients underwent targeted fusion biopsy followed by concomitant standard 12-core double sextant PBx of the peripheral zone. Using the UroNav MRI-US fusion system, targeted fusion PBx was performed with the previously identified mpMRI lesions superimposed on the real-time TRUS images. Each lesion was sampled in the transverse or sagittal plane by an end or side fire TRUS probe with a minimum of 2 cores per target lesion. After completing targeted fusion PBx the targeting system was turned off and a standard 12-core PBx in a double sextant template was performed by the same urologist who performed the targeted component of the biopsy.
In patients on active surveillance of low risk prostate cancer 2 to 4 anteriorly directed cores were also periodically obtained, which were included in our analysis. Only TRUS images were used for the standard template component of PBx. Three urological oncologists, each with more than 10 years of transrectal prostate biopsy experience, performed all biopsies.
Patient characteristics included in study were age, prebiopsy PSA, digital rectal examination result, mpMRI indication and prostate volume on mpMRI.
Regarding mpMRI characteristics, the number of target lesions, target lesion size and location were recorded. Location (peripheral zone vs transitional zone vs anterior fibromusclar stroma) was described for each target lesions. Each target lesion was graded according to PI-RADS version 2 by one of our dedicated genitourinary radiologists with more than 100 prostate MRI readings each. Studies read prior to PI-RADS version 2 and those with missing PI-RADS scores were rereviewed by one of us (AM or MH). 10Biopsy cores were evaluated individually by a dedicated genitourinary pathologist.
The study primary end point was the detection of any PCa. The secondary end point was the detection of csPCa, defined as GG 2 or greater prostate adenocarcinoma.
Descriptive statistics were used to describe patient characteristics. The McNemar test was applied to compare the proportion of overall cancer detection and csPCa detection among standard biopsy vs targeted fusion PBx results in the overall cohort. The added benefit was measured by the percent of csPCa missed if only 1 technique had been used. Subgroup analysis was done according to the mpMRI indication, the highest PI-RADS score, PSA density (less than 0.15 vs 0.15 ng/ml/cc or greater), largest lesion size (10 or less vs greater than 10 mm), location of the dominant mpMRI lesion and the number of reported targets. All statistical analyses were performed using STATA® 13 with 2-sided statistical significance considered at p <0.05.
A total of 506 patients with identifiable target lesions on mpMRI were referred for fusion PBx between March 2014 and January 2018 ( table 1). Mean patient age was 65.6 years, median PSA was 7.0 ng/ml and median prostate volume was 48 cc. Indications included elevated PSA with a previous negative PBx in 46% of cases, PCa while on active surveillance in 35%, elevated PSA without a prior PBx in 15% and an isolated abnormal digital rectal examination in 3%. Of the patients 72% had only 1 identifiable target lesion on mpMRI while 24% and 4% had 2 and 3 targets, respectively. The median size of all target lesions was 12 mm.
Overall 670 target lesions were biopsied and PI-RADS 4 was identified in 46%. Figure 1 shows pathological results by the PI-RADS score. The overall detection rate was 12%, 19%, 51% and 72% for PI-RADS 2, 3, 4 and 5, and the csPCa detection rate was 8%, 11%, 26% and 45%, respectively ( fig. 1).
The overall PCa detection rate was 57.7% and 54.0% for standard and fusion PBx, respectively (p=0.12, table 2). For csPCa the detection rate was 30.8% and 24.7% for fusion and standard template PBx, respectively (p=0.001). The difference in csPCa detection between fusion and standard template PBx was statistically significant in patients with elevated PSA and a previous negative PBx (30.6% vs 23.4%, p=0.01), those with high PI-RADS score target lesions (PI-RADS version 2 score 4, 31.0% vs 24.1% and score 5, 50.4% vs 39.3%, each p=0.01), those with a larger than 10 mm target lesion on mpMRI (34.2% vs 26.3%, p=0.001), those with PSA density greater than 0.15 ng/ml/cc (46.1% vs 34.0%, p <0.001) and those with only 1 target lesion reported on mpMRI (29.5% vs 22.9%, p=0.001).
Figure 2 shows the proportion of prostate cancer cases detected by each biopsy modality according to Gleason GG. Of 348 patients with any PCa the lesion would have been missed in 75 (22%) by fusion PBx alone and in 56 (16%) by standard PBx alone ( fig. 2, A). Furthermore, in 29 of 185 patients (16%) diagnosed with csPCa the lesion would have been missed if only fusion PBx had been done (an added benefit of standard PBx) and the lesion would have been missed in 60 (32%) using standard PBx alone (an added benefit of fusion PBx). Figure 2, Bshows the percent of csPCa missed by fusion biopsy according to the mpMRI indication. The supplementary table ( https://www.jurology.com) lists the added benefit of each technique in all subgroups. Table 3 shows the pathology results of patients according to biopsy technique.
The exact role of mpMRI in the initial diagnosis and management of PCa has still not been clearly determined. 7Siddiqui et al found that mpMRI and fusion PBx significantly increased the csPCa detection rate while decreasing the detection of low grade, clinically insignificant PCa. 11The reported negative predictive value of mpMRI for overall PCa and csPCa widely varies in the literature. 12The PROMIS (PROstate MR Imaging Study) study showed a 76% negative predictive value (range 69% to 82%) for GG 2 or greater PCa using mpMRI against template mapping PBx. 13This suggests that targeted biopsy alone can miss a significant proportion of csPCa.
The PRECISION (Prospective Randomized Evaluation of Celecoxib Integrated Safety versus Ibuprofen or Naproxen) trial demonstrated that MRI targeted biopsy alone without systematic PBx identified more csPCa than standard systematic PBx while also more frequently avoiding the detection of clinically insignificant disease. 14However, it is unknown how many csPCa cases were missed in the mpMRI group. Conversely the ASIST (Active Surveillance magnetic resonance Imaging Study) trial, in which grade group 1 PCa was studied in men on active surveillance, MRI targeted biopsy did not result in a significant difference in PCa upgrading compared with systematic biopsy alone. 15
Thus, the debate is still ongoing regarding the possibility of not performing standard PBx and only performing targeted fusion PBx when a suspicious lesion is identified on mpMRI. This could potentially enable a reduction in the complication rate and in patient discomfort as well as decrease even further the diagnosis of low risk PCa. It is our current practice to perform fusion as well as standard PBx in all patients found to have a target lesion on mpMRI. We investigated the added value of standard PBx to targeted fusion PBx or more accurately the risk of omitting the random standard template.
Several findings are notable. 1) We noted similar overall cancer detection rates for standard and fusion PBx (57.7% and 54.0%, respectively). However, fusion PBx was associated with a 25% higher detection rate of csPCa (30.8% vs 24.7%). These values are in line with a report by Siddiqui et al, who found 16% higher detection of csPCa according to our definition of csPCa (31.3% vs 26.2%). 11
The cancer detection rate of mpMRI targeted PBx varies greatly according to the clinical indication. Presumably the patient with a prior negative standard template PBx will be more likely to have csPCa detected on targeted biopsy vs repeat standard PBx. This was seen in our results as the difference in the csPCa detection rate was larger in patients with a prior negative PBx. 16Thus, it is important to consider patient selection and the indication when evaluating differences in cancer detection techniques.
2) The difference in cancer detection varied by radiological and clinical characteristics. The csPCa detection rate increased progressively with higher PI-RADS scores and fusion PBx demonstrated greater improvement than standard PBx in cases of higher graded lesions. We found that fusion PBx performed better than standard template PBx in patients with larger lesions, those with high PSA density (greater than 0.15 ng/ml/cc) and those with only 1 reported lesion on mpMRI. That being said, with the increasing use of mpMRI the accuracy and detection rate of targeted biopsy may be increased by improving imaging quality and standardizing radiological training and reporting.
3) We found that approximately 1 of 5 patients with overall PCa and 1 of 6 with csPCa found by standard and/or fusion PBx would be missed if only targeted fusion PBx were performed. These numbers are similar to those in other large series at high volume centers using similar biopsy protocols, which show that between 13% and 21% of csPCa cases would be missed if only targeted fusion PBx were peformed. 11,17,18Differences can be explained by different cohort compositions with regard to the relative percent of patients with different mpMRI clinical indications. Indeed, patients with proven PCa on prior biopsy had a higher rate of missed csPCa on targeted fusion PBx alone than patients with a prior negative PBx and those with no prior PBx (19% vs 14% and 12%, respectively).
That specific subgroup was examined in the recently published MRI-FIRST trial (Assessment of Prostate MRI Before Prostate Biopsies). 19The investigators concluded that standard and targeted biopsy should be performed and those numbers are comparable to ours. Of 206 biopsy naïve patients who underwent standard as well as targeted biopsy 89 were diagnosed with csPCa, of whom 19 (21%) would have been missed if only targeted biopsy had been done.
Potential mechanisms of fusion PBx failure include radiologist inexperience, invisible mpMRI PCa, fusion PBx technique error and intralesion Gleason heterogeneity. To our knowledge the clinical implications of csPCa undetected by fusion PBx remain unknown.
Altogether our results suggest that fusion PBx alone is associated with a higher detection rate of csPCa and a lower detection rate of clinically insignificant PCa. Future studies are needed to determine whether the increase in csPCa detection using fusion PBx would optimize the initial selection classification of patients who are candidates for treatment or whether it would increase the detection and treatment of what we label csPCa which may not have a worse outcome in the long term.
Ultimately this will depend on how much a clinician is willing to accept a trade-off between fewer biopsy cores taken and some csPCa missed. In our opinion there is an added benefit to performing standard template biopsy, in addition to targeted fusion PBx, in patients with 1 or more mpMRI targets. In our cohort standard template biopsy provided the most added benefit in patients with PI-RADS 3 lesions with 45% of csPCa missed if only targeted biopsy was done (supplementary table, https://www.jurology.com). It is possible that, for example, in a patient with a prior negative biopsy, high PSA density and a large high grade PI-RADS lesion standard template biopsy might be omitted with a smaller chance of missing csPCa. All of these conditions are associated with a smaller added benefit from standard biopsy. Accordingly, with more data in future analyses the added usefulness of standard template PBx may ultimately be individualized and vary by clinical indication, radiographic findings and biochemical results.
Our study is not without limitations. 1) We report our single institution retrospective experience.
2) Our patient population was heterogeneous with regard to biopsy indication, although subgroup analyses were performed. During the study period mpMRI protocols were not homogeneous in all patients, especially in regard to magnet strength (1.5 vs 3.0 Tesla) or use of an endorectal coil. However, to our knowledge it is not established that the detection rate significantly differs between examinations performed at 1.5 Tesla with an endorectal coil vs 3.0 Tesla without an endorectal coil. 20
3) During the procedures a single operator performed fusion PBx as well as standard template PBx. Although the software registration was turned off during template PBx, it is possible that some form of cognitive fusion influenced the exact location of these biopsies.
4) These results are based on mpMRI readings and biopsy histological grading by dedicated genitourinary radiologists and genitourinary pathologists, respectively. These results may not be applicable among less subspecialized practitioners with less experience.
5) Our definition of csPCa might not encompass all truly significant disease since high volume Gleason 6 might represent meaningful disease and low volume Gleason 7 (3 + 4) may also be included in non-csPCa.
6) In the current report we did not rely on whole mount pathological analysis, which precluded knowledge of the true tumor burden.
Fusion PBx is associated with a higher csPCa detection rate but it would miss a significant proportion of csPCa if performed alone without the addition of standard template PBx. Therefore, standard template PBx should still be included in all fusion biopsy programs. However, the added value may be lower in patients who already had a negative PBx or when targeted biopsy is more likely to be positive based on higher PSA density, high grade lesions and/or larger targets.
1. Tooker GM, Truong H, Pinto PA et al: National survey of patterns employing targeted MRI/US guided prostate biopsy in the diagnosis and staging of prostate cancer. Curr Urol 2019; 12: 97