Multiparametric MRI vs. HistoScanning™ For Prostate Cancer Detection
By: Dan Sperling, MD
Until the recent past, the traditional clinical pathway for diagnosing prostate cancer (PCa) moved quickly from a patient’s abnormal PSA/DRE to a transrectal needle biopsy of the prostate guided by ultrasound (TRUS biopsy). Much has been written about the problems with this random blind approach; its flaws include overdetection of insignificant PCa, underdetection of significant PCa, false negative results, pain, risks of bleeding, infection and urinary/sexual complications. Today, multiparametric MRI (mpMRI) is finding increasing favor as a powerful tool for detecting prostate cancer because it reveals tissue differences whereas ultrasound does not. The new diagnostic pathway includes mpMRI scanning before a biopsy in order to determine if suspicious areas are present, and if a biopsy is warranted, a targeted biopsy can be conducted under real-time MRI guidance. This pathway offers two main advantages:
- mpMRI can rule out the need for a biopsy yet continue to be used to monitor if needed, and
- If a biopsy proves necessary, a targeted biopsy reduces the number of needles used.
From the perspective of urologists, who predominantly manage PCa patients, integrating MRI into their patient evaluations poses various challenges. Urologists have ultrasound equipment placed in their office practices whereas patients must be referred to an imaging radiologist at an outside facility; most urologists are not trained to interpret MRI scans; the technology to fuse MRI scans with real-time ultrasound is expensive and involves a learning curve, plus the patient must still be referred out for an MRI. Therefore, a portable ultrasound-based imaging device that would accurately distinguish between a prostate cancer tumor and healthy tissue would be a sort of Holy Grail for urologists.
In January, 2008 a new technology called HistoScanning™ was introduced. Described as “a proprietary tissue differentiation, visualization and measurement technology designed to assist specialists in identifying changes to solid organ tissues”[i], the software integrates with the transrectal ultrasound with which urologists are already familiar. Thus, there was good reason for optimism that here was a breakthrough that could achieve the same purpose as mpMRI. Sadly, that hope has not been borne out.
Before comparing how HistoScanning™ compares with mpMRI of the prostate, here are observations from recent published studies on HistoScanning™ performance:
- Schiffmann et al. (2015)[ii] tested HistoScanning™ at three different signal volume cutoffs on each of 40 patients suspected of PCa. Based on areas identified by the software as possible tumors, targeted biopsies were conducted as well as standard TRUS biopsies. The biopsy results were compared, and the researchers found the HistoScanning™ prediction of cancerous areas unreliable at all signal levels. They concluded, “We cannot recommend a variation of well-established biopsy standards or reduction in biopsy cores based onHistoScanning™ signals.”
- Hamann et al. (2015)[iii] used HistoScanning™ with 97 patients who had previous negative biopsies to determine if the technology could detect PCa. They conducted targeted biopsies into the suspicious areas identified by HistoScanning™ followed by 14-needle standard biopsies. The authors found that sensitivity (45%) and positive predictive value (19%) were low, and cautiously concluded that the information from HistoScanning™ “facilitates [biopsy] targeting to vulnerable foci and results in a higher cancer detection rate.” They also recommended that some form of imagining be used prior to biopsy to improve detection.
- Schiffmann et al. (2015)[iv] conducted a review of published articles on HistoScanning™ since 2008, and found the data inconsistent and therefore controversial. They analyzed 12 studies along three lines: prediction of final pathology at radical prostatectomy, prediction of disease stage and application atprostate Initially, a pilot study demonstrated high sensitivity and specificity when predicted disease sites matched those found post-prostatectomy. However, subsequent studies based on the same criteria did not demonstrate the same results. In addition, staging information was unreliable especially with regard to tumor extension outside the prostate capsule and seminal vesicle invasion. In large studies, HistoScanning™ performed poorly in predicting positive biopsies. In fact, standard TRUS biopsies performed better than those augmented by HistoScanning™. According to the team, “Currently, evidence is at best weak and questions whether HistoScanning™ might improve the detection of PCa.”
Comparisons between mpMRI and HistoScanning™ are still scant. However, the existing evidence suggests that mpMRI is the superior of the two. In 2013, a research group out of University College London announced implementation of a well-designed comparison study called the PICTURE study. For each study participant, mpMRI, HistoScanning™ and transperineal mapping biopsy would be conducted; the biopsy results would be the standard for evaluating the accuracy of both imaging types. Although final data has not been published, early results are clear. An interim analysis was announced in 2014 at the European Association of Urology 29th Annual Congress: “‘We think that mpMRI might be a useful test at this point in the diagnostic pathway, and its encouraging negative predictive value may allow men to safely avoid biopsy,’ said lead researcher Lucy Simmons, MD.” However, HistoScanning “‘did not perform well at all for this…’”[v] The session chair, Dr. Boris Hadaschick, commented, “”HistoScanning is of very limited value… I think recent publications show that it’s not very good.”
Most recently, Orcqyk at el. (2016)[vi] published a prospective, blinded comparative analysis between mpMRI and HistoScanning™ among 31 PCa patients scheduled for prostatectomy. The results of each patient’s mpMRI, HistoScanning™ and pathology of post-surgery prostate specimens were read by a single radiologist, urologist and pathologist, respectively, in blinded fashion. The PCa location reports generated by each clinician were then compared and evaluated for concordance by a committee composed of 2 urologists, 2 pathologists, and 1 radiologist. HistoScanning™ identified PCa correctly in 46.2% suspicious areas compared with mpMRI’s 52.6%. Overall, mpMRI had significantly better accuracy, positive predictive value, negative predictive value and specificity. The authors found that, “Among tumors with Gleason score>6, mpMRI detected 19/22 (86.4%) whereas HS detected only 11/22 (50%). Similarly, among tumors>10mm in maximal diameter, mpMRI detected 28/34 (82.4%) whereas HS detected only 19/34 (55.9%).” They concluded that HistoScanning™ is inferior to mpMRI in performance, and the technology needs further refinement before having wide application.
Until further technical developments in ultrasound, mpMRI is expected to remain the imaging modality of choice for the detection of prostate cancer.
[ii] Schiffmann J, Mehring G, Tennstedt P, Manka L et al. True targeting-derived prostate biopsy: HistoScanning™ remained inadequate despite advanced technical efforts. World J Urol. 2015 Jul 28. [Epub ahead of print]
[iii] Hamann MF, Hamann C, Trettel A, Jünemann KP, Naumann CM. Computer-aided transrectal ultrasound: does prostate HistoScanning™ improve detection performance ofprostate cancer in repeat biopsies? BMC Urol. 2015 Jul 30;15:76.
[iv] Schiffmann J, Manka L, Boehm K, Leyh-Bannurah SR Controversial evidence for the use of HistoScanning™ in the detection of prostate cancer. World J Urol. 2015 Dec;33(12):1993-9.
[v] Kate Johnson. “Can Imaging Replace Biopsy For Some Prostate Cancer?” Medscape Medical News. April 24, 2014.
[vi] Orcqyk C, Rosenkrantz AB, Deng FM, Melamed J et al. A prospective comparative analysis of the accuracy of HistoScanning and multiparametric magnetic resonance imaging in the localization of prostate cancer among men undergoing radical prostatectomy. Urol Oncol. 2016 Jan;34(1):3.e1-8.