Ruling out clinically significant prostate cancer with negative multi-parametric MRI

Abstract

Purpose

To evaluate the negative predictive value (NPV) of a negative prostate multi-parametric magnetic resonance imaging (mpMRI) in ruling out clinically significant prostate upon 12-core systematic biopsy.

Methods

We retrospectively reviewed 114 men evaluated at our institution who underwent systematic 12-core biopsy within 1 year of a negative prostate mpMRI. Clinicopathologic features were evaluated and NPV was calculated for detection of clin-ically significant (Gleason ≥ 7) cancer. Regression analysis was performed to identify clinical predictors of biopsy outcome.

Results

Overall, 88 (77.2%) patients in our cohort had no cancer detected upon biopsy. The highest pathologic grade was Gleason 6 (3 + 3) in 22 (19.3%) patients, and Gleason ≥ 7 in 4 (3.6%) patients. NPV for detecting Gleason ≥ 7 cancer was 96.5% (95% CI 93.1–99.9%) in the entire negative MRI cohort, 100% in those who were prostate biopsy naive (n = 20), 100% in those with a prior negative biopsy (n = 53), and 90% in those who have had a previous positive biopsy and on active surveillance (n = 41). Regression analysis identified no predictors of significant cancer in our cohort.

Conclusion

In our cohort of men with no lesions detected on prostate mpMRI, we found very low rates of clinically significant cancer on systematic 12-core biopsy. In the few patients who diagnosed with prostate cancer, the majority had low-risk disease and could remain on active surveillance. Although validation studies and greater sample size is needed before clinical recommendations can be made, our data suggest patients with negative mpMRI evaluated by experienced radiologists may avoid unnecessary prostate biopsy and potential overtreatment.

Introduction

Prostate multi-parametric MRI (mpMRI) has significantly changed the paradigm of prostate cancer evaluation. The high anatomic resolution of T2 weighted imaging (T2WI) and simultaneous functional sequences (apparent diffusion coefficient [ADC] maps, diffusion weighted imaging [DWI] with high b value, and dynamic contrast-enhanced [DCE] imaging) provide superior visualization of prostate tissue and identification of clinically significant (Gleason ≥ 7) lesions [1-3], A “negative” magnetic resonance imaging (MRI) without any distinct regions of suspicion is assigned as Prostate Imaging—Reporting and Data System (PI-RADS) category 1 lesion based on PI-RADS version 2 (v.2) guidelines [1-4]. Based on these guidelines, patients with PI-RADS category 1 imaging should have a “very low” likelihood of having any clinically significant PCa; however, these patients with a negative MRI still undergo systematic TRUS-guided biopsies due to possibility of having magnetic resonance (MR) invisible cancers. Why did you change to MR invisible here? Why not mpMRI?

We hypothesize PI-RADS category 1 mpMRI be used an imaging-based biomarker to rule out clinically significant PCa. The ability to rule out high-risk PCa with this non-invasive imaging modality could allow select patients to avoid or delay unnecessary procedures. The risk and cost of biopsy associated adverse events are not insignificant. Although contemporary rates of infection are only about 3.5–4%, rates of hospitalization are increasing [5, 6]. The multicenter prospective PROMIS trial evaluating mpMRI as a triage test supported of the use of mpMRI; however, this study may not be applicable to the general population and may be limited in addressing the predictive value of a negative mpMRI [7]. The PROMIS trial used a limited study population (only biopsy naïve patients, no patients with prostate volume > 100 ml), 1.5T scanner without use of endorectal coil, and negative MRI imaging (defined on a Likert Scale) included low suspicious ion lesions. Therefore, it has limited application to answering our current study question regarding completely negative mpMRI.

Current data on negative prostate MRI is scarce, but of great clinical interest. In our present study, we aim to evaluate a contemporary cohort of patients without any lesions identified on prostate mpMRI to assess the utility of systemic prostate biopsy in presence of a negative mpMRI study.

Materials and methods

Study population

We retrospectively reviewed 114 men evaluated at our institution who underwent biopsy within 1 year of having a negative prostate mpMRI. mpMRI were obtained between January 2013 and February 2017 under institutional review board-approved protocols for retrospective review of patients imaged at the National Institutes of Health. Patients with previous treatment for PCa (surgical, androgen deprivation, immunotherapy, and radiation therapy) were excluded. For patients with multiple negative mpMRI and biopsy sessions, only initial imaging and biopsy results were considered. The decision to undergo biopsy was ultimately determined by the patient based on their preference after review of MRI results. Patients with negative mpMRI who did not undergo subsequent biopsy at our institution were reviewed for details regarding outside procedures within 1 year of their mpMRI.

Imaging and biopsy protocols

A 3.0 Tesla (3T) MRI (Achieva, Philips Medical Systems, Best, Netherlands) was used to acquire all images. A previously published prostate mpMRI protocol was followed to obtain T2 weighted imaging (T2WI) with apparent diffusion coefficient (ADC) mapping, diffusion weighted imaging (DWI) with high b value (b = 2000 s/mm²), and dynamic contrast-enhanced (DCE) sequences. A 16-channel surface-array coil (SENSE cardiac surface coil, Philips Medical Systems, Best, the Netherlands) positioned over the pelvis and an endorectal coil (BPX-30, BAYER, Medrad, Pittsburgh, Pennsylvania) were also utilized. A single genitourinary radiologist with 9 years of experience interpreting prostate mpMRI evaluated all images.

All patients included in this analysis underwent standard-of-care, non-targeted transrectal ultrasound (TRUS) guided systematic 12-core biopsy in an outpatient setting. A single, genitourinary pathologist reviewed all core specimens. Clinically significant disease was defined as having any Gleason ≥ 7 (3 + 4) on histologic evaluation, and the highest histologic grade was recorded for each patient’s biopsy results. For patients who did not undergo systemic biopsy at National Institutes of Health, tissue from outside institutions was internally reviewed. Verified outside pathology reports were considered when specimens were not available.

Data analysis

Baseline clinical information was evaluated using the appropriate descriptive statistics. Subgroup analysis based on institutional versus outside biopsy accruement was compared using appropriate comparative statistical tests (Rank-Sum test for continuous variables, Fisher’s Exact test for categorical variables). Pathological outcomes and negative predictive value (NPV) for ruling out clinically significant PCa were reported for the entire cohort and for cohorts by biopsy history (biopsy naïve, prior negative, and prior positive on active surveillance).

Univariate and multivariate logistic regression analyses were performed to identify predictors of PCa detection in this cohort. Variables of interest for the logistic regression analysis were age, race (African American vs. non-African American), PSA, prostate volume, and the presence of > cT1c disease (positive DRE). All statistical analysis was performed using Stata 14 (StataCorps LLC, College Station, Texas) with statistical significance considered at 2-sided p < 0.05.

Results

Baseline demographics

Of the 114 men included in our analysis, 25 (21.9%) had biopsy at outside institutions, while 89 (78.1%) had biopsy at our institution (Table 1), These two cohorts displayed similar baseline characteristics aside from the time between mpMRI and biopsy (− 4 vs. 2 months, p < 0.001). Median age and PSA for the entire cohort was 61 (IQR 57–67) years and 5.5 (IQR 3.6-8.7) ng/ml for the entire cohort, respectively. Twenty (17.5%) patients were biopsy naïve, 53 (46.5%) reported only prior negative biopsies, and 41 (36.0%) reported prior positive biopsy and were on active surveillance at the time of mpMRI.

Table 1.

Baseline demographic of patients with negative prostate mpMRI undergoing systematic ultrasound guided biopsy

	Overall	Bx before mpMRI	Bx after mpMRI	p value, between Bx before versus after mpMRI
Patients [n (%)]	114	25 (21.9%)	89 (78.1%)
Age (years)	61 (57–67)	61 (57–66)	62 (57–68)	0.24
Ethnicity [n (%)]
Caucasian	93 (81.6)	20 (80.0)	73 (82.0)	0.79
African American**	12 (10.5)	4 (16.0)	8 (9.0)
Hispanic**	2 (1.8)	0 (0.0)	2 (2.3)
Unknown**	7 (6.1)	1 (4.0)	6 (6.7)
PSA, ng/ml	5.5 (3.6–8.7)	5.7 (3.8–10.1)	5.5 (3.5–8.1)	0.56
Prostate volume, ml	57 (42–80)	59.5 (39–82.0)	57 (43–80)	0.55
> cT1c disease [n (%)]	5 (4.4)	0 (0.0)	5 (5.6)	0.58
Time from mpMRI to biopsy (months)	1 (0–3)	− 4 (− 7 to − 3)	2 (0–4)	< 0.001*
Prior positive biopsy [n (%)]	41 (36.0)	6 (24.0)	35 (39.3)	0.1
6 (3 + 3) [n (%)]	37 (90.2)	5 (83.3)	32 (91.4)	0.48
7 (3 + 4) [n (%)]***	4 (9.8)	1 (16.7)	3 (8.6)

Values are expressed as medians (interquartile range) unless otherwise specified

Bx biopsy, PSA prostate specific antigen, mpMRI multi-parametric magnetic resonance imaging

^*Statistical significance

^***omitted due to low incidence

Biopsy results

Pathology outcomes are presented in Table 2 and Fig. 1. A total of 26 (23%) patients had cancer detected on systematic TRUS biopsy. The majority (22 men, 19.3%) had Gleason 6 (3 + 3) disease. Clinically, significant cancer was detected in four (3.6%) men, all of whom had prior positive biopsies and were on active surveillance (NPV = 90.0%) No clinically significant cancer was detected in biopsy naïve patients (NPV = 100.0%) and in patients who had prior negative biopsies (NPV = 100.0%).

Table 2.

Pathology results from patients with negative prostate mpMRI, stratified based on prior biopsy history

	Overall, n = 114	No prior biopsy, n = 20	Prior biopsy, only negative, n = 53	Prior biopsy, positive, n = 41
No cancer [n (%)]	88 (77.2)	15 (75.0)	50 (94.3)	23 (56.1)
Gleason 6 (3 + 3) [n (%)]	22 (19.3)	3 (5.7)	3 (5.6)	14 (34.2)
Gleason 7 (3 + 4) [n (%)]	2 (1.8)	0 (0.0)	0 (0)	2 (4.9)
Gleason 7 (4 + 3) [n (%)]	2 (1.8)	0 (0)	0 (0)	2 (4.9)
NPV for any cancer (%)	77.2	75.0	94.3	56.1
NPV for significant cancer (%)	96.5	100.0	100.0	90.0

Gleason ≥ 7 (3 + 4) is considered clinically significant cancer. Prior positive cohort included those who had a prior positive biopsy > 1 year prior to their negative mpMRI

NPV negative predictive value

Fig. 1.

Pathology results from systematic 12-core biopsies of patients with negative prostate MRI

Predictive associations

Neither univariate nor multivariate logistic regression analysis identified any clinical parameters predictive of PCa in this negative MRI cohort (Table 3).

Table 3.

Associations of patient parameters with detection of any prostate cancer

	Univariate		Multivariate
	OR (95% CI)	p value	OR (95% CI)	p value
Age	0.99 (0.93–1.05)	0.79	1.01 (0.95–1.07)	0.82
Race (AA vs. not AA)	1.14 (0.29–4.58)	0.85	0.98 (0.23–4.19)	0.98
PSA	0.90 (0.80–1.01)	0.19	0.99 (0.81–1.21)	0.93
Prostate volume	0.99 (0.97–1.00)	0.06	0.98 (0.96–1.01)	0.19
> cT1c disease	2.36 (0.37–14.95)	0.36	1.41 (0.21–9.55)	0.73

OR odds ratio, CI confidence interval, AA African American, PSA prostate specific antigen, mpMRI multi-parametric magnetic resonance imaging, Bx biopsy

^*Omitted due to low incidence

Discussion

While there is excellent data on missed cancers in patients with MR visible lesions who undergo both targeted and systematic biopsy [1, 3, 8, 9], less is known regarding patients with negative mpMRI who may not undergo biopsy. Despite optimized image acquisition and expert interpretation, significant cancers may still be missed on imaging. Understanding the likelihood of significant cancer despite a negative mpMRI has both scientific merits as well as potential to inform clinical decision-making.

Of the 114 men with negative mpMRI who had undergone biopsy, only four (3.6%) were found to have clinically significant cancer. All four patients had previously diagnosed Gleason 6 (3 + 3) PCa. While a negative mpMRI cannot unequivocally rule out significant PCa, the likelihood of finding significant PCa identified in patients with negative mpMRIs is low (NPV = 96.5%). Our results are similar to other retrospective analyses evaluating the NPV of negative prostate mpMRI; Wysock et al. and Lu et al. reported NPV of 98.7 and 97.0%, respectively [10, 11]. We previously reported mpMRI assigned “low-risk” using the NIH scoring scale is unlikely to harbor significant prostate cancer, with an NPV of 92.0%. However, this study was from the early-prostate MRI era prior to standardized guidelines and some patients had lesions identified of low-suspicion [12]. Filson et al. performed a similar analysis but reported a NPV of 84% in their cohort of 113 men and concluded that men with clinical suspicion of PCa should undergo biopsy even with negative mpMRI [13]. The etiology of the lower NPV is unclear; however, their imaging protocol and study population greatly differed from our study: no use of endorectal coil, inclusion of early-MRI era patients from 2009 to 2013, use of template based 12-core biopsy, with a greater proportion of patients with previous positive biopsy (47.5%). A comparison of these contemporary studies evaluating negative mpMRI and subsequent pathologic outcomes is presented in Table 4.

Table 4.

Comparison of studies evaluating NPV of negative/no lesion mpMRI for detection of clinically significant prostate cancer

	Overall	No prior biopsy	Prior biopsy, only negative	Prior biopsy, positive
An et al. (present study) (n = 114)*	96.5%	100.0%	100.0%	90.0%
Lu et al. [11] (n = 100)	97.0%	100.0%	96.0%	93.1%
Boesen et al. [14] (n = 100)	N/A	N/A	95.0%	N/A
Wysock et al. [10] (n = 75)	98.7%	100.0%	96.0%	100.0%
Filson et al. [13] (n = 244)	84.0%	N/A	N/A	N/A
Itatani et al. [15] (n = 193)**	89.6%	N/A	N/A	N/A
Yerram et al. [12] (n = 125)	92.0%	N/A	N/A	N/A

NPV negative predictive value

Gleason ≥ 7 (3 + 4) is considered clinically significant cancer

^*denotes current study

^**denotes use of 1.5T MRI scanner versus 3T MRI in other respective studies; Lu et al. [11], Boesen et al. [14], Wysock et al. [10]. Filson et al. [13], Itatani et al. [15] and Yerram et al. [12]

Our findings may be particularly useful for patients who have had previous negative biopsy but are unsure of its reliability. None of the 53 patients with prior negative biopsy had any clinically significant cancer. Of the three patients within this group who did have cancer, all had low-risk Gleason 6 (3 + 3) disease. Boesen et al. had similar findings in their cohort of 194 men with low-suspicion mpMRI and prior negative biopsy (NPV 95%) [14]. Our combined evidence suggests repeat biopsy in those patients with negative mpMRI and prior negative biopsy may be of limited value.

There are several limitations to our study. First, generalizability is limited due to patient homogeneity and small sample size. Our patients were all evaluated at a research institution with high-volume and experience specialists. Our results may be superior than what would be observed in the community. Furthermore, our sample size of 114 men is relatively small and only 20 patients were biopsy naïve. Another limitation is some of the negative mpMRI patients did not undergo biopsy. Although patient preference was the main determining factor for whether a biopsy was performed, bias from clinician counseling and a myriad of other social and clinical factors likely contributed to the decision. Furthermore, our results likely underestimate the true presence and grade of cancer since a systematic 12-core biopsy was used as the surrogate end point. Current literature reports discordance of pathology between 12-core biopsy and radical prostatectomy to be as high as 30–40% [9]. However, whole gland pathology was not available because radical prostatectomy was not indicated for our patients. The long natural history of prostate cancer and growing use of mpMRI in different practice settings (both academic and community practices) invites future studies with longer follow-up, greater sample size, and more heterogeneous cohorts. Despite the inherent limitations of our study, our findings contribute to the growing understanding of MRI technology for the evaluation of PCa.

Conclusion

In our cohort of men with no lesions detected on prostate mpMRI, we found very low rates of clinically significant cancer on systematic 12-core biopsy. In the few patients who diagnosed with prostate cancer, the majority had low-risk disease and could remain on active surveillance. Although validation studies and greater sample size is needed before clinical recommendations can be made, our data suggest patients with negative mpMRI evaluated by experienced radiologists may avoid unnecessary prostate biopsy and potential overtreatment.