DWI and PRECISE criteria in men on active surveillance for prostate cancer: A multicentre preliminary experience of different ADC calculations

DWI and PRECISE criteria in men on active surveillance for prostate cancer: A multicentre preliminary experience of different ADC calculations

Journal Pre-proof DWI and PRECISE criteria in men on active surveillance for prostate cancer: A multicentre preliminary experience of different ADC ca...

5MB Sizes 0 Downloads 19 Views

Journal Pre-proof DWI and PRECISE criteria in men on active surveillance for prostate cancer: A multicentre preliminary experience of different ADC calculations

Francesco Giganti, Martina Pecoraro, Davide Fierro, Riccardo Campa, Francesco Del Giudice, Shonit Punwani, Alex Kirkham, Clare Allen, Mark Emberton, Carlo Catalano, Caroline M. Moore, Valeria Panebianco PII:

S0730-725X(19)30574-0

DOI:

https://doi.org/10.1016/j.mri.2019.12.007

Reference:

MRI 9357

To appear in:

Magnetic Resonance Imaging

Received date:

20 September 2019

Revised date:

11 December 2019

Accepted date:

27 December 2019

Please cite this article as: F. Giganti, M. Pecoraro, D. Fierro, et al., DWI and PRECISE criteria in men on active surveillance for prostate cancer: A multicentre preliminary experience of different ADC calculations, Magnetic Resonance Imaging(2018), https://doi.org/10.1016/j.mri.2019.12.007

This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

© 2018 Published by Elsevier.

Journal Pre-proof DWI and PRECISE criteria in men on active surveillance for prostate cancer: a multicentre preliminary experience of different ADC calculations Francesco Giganti a,b,c, Martina Pecoraro c, Davide Fierro c, Riccardo Campa c, Francesco Del Giudice e, Shonit Punwani a,d, Alex Kirkham a, Clare Allen a, Mark Emberton b,f, Carlo Catalano c, Caroline M Moore b,f*, Valeria Panebianco c*

ro of

* These authors share joint senior authorship

a. Department of Radiology, University College London Hospital NHS Foundation Trust, London, UK

-p

b. Division of Surgery & Interventional Science, University College London, London,

re

UK

c. Department of Radiological Sciences, Oncology and Pathology, Sapienza

lP

University of Rome, Rome, Italy

d. Centre for Medical Imaging, University College London, London, UK e. Department of Urology, Sapienza University of Rome, Rome, Italy Department of Urology, University College London Hospital NHS

na

f.

Jo

ur

Foundation Trust, London, UK

Corresponding author:

Francesco Giganti, MD Division of Surgery and Interventional Science, University College London, 3rd Floor, Charles Bell House, 43-45 Foley St., London, United Kingdom W1W 7TS email: [email protected]

Journal Pre-proof DWI and PRECISE criteria in men on active surveillance for prostate cancer: a multicentre preliminary experience of different

Jo

ur

na

lP

re

-p

ro of

ADC calculations

Journal Pre-proof Abstract Purpose: The PRECISE score estimates the likelihood of radiological progression in patients on active surveillance (AS) for prostate cancer (PCa) with serial multiparametric magnetic resonance imaging (mpMRI). A PRECISE score of 1 or 2 denotes radiological regression, PRECISE 3 indicates stability and PRECISE 4 or 5 implies progression. We evaluated the inter-reader reproducibility of different apparent diffusion coefficient (ADC) calculations and their relationship to the

ro of

PRECISE score.

Material and methods: Baseline and follow-up scans (on the same MR systems) of 30 patients with visible lesions from two different institutions

-p

were analysed by two radiologists. The PRECISE score was initially

re

assessed in consensus. At least six weeks later, to reduce the likelihood of being influenced by the consensus PRECISE reading, each radiologist

lP

independently calculated ADC for the following: lesion, non-cancerous tissue and urine in the bladder. Normalised ADC ratios were calculated

na

with respect to normal prostatic tissue (npADC) and urine. Spearman’s correlation (ρ), intraclass correlation coefficients (ICC), differences in ADC

ur

and ROC curves were computed.

Jo

Results: Interobserver reproducibility was very good (ρ>0.8; ICC>0.90). Lesion ADC (0.91 vs 0.73 × 10−3 mm2/s; p=0.025) and npADC ratio (0.68 vs 0.53; p=0.012) at follow-up mpMRI were different between patients with radiological regression or stability vs progression. Cut-offs of 0.77 × 10−3 mm2/s (lesion ADC) and 0.59 (npADC ratio) could differentiate the two groups (area under the curve: 0.74 and 0.77, respectively). Conclusion: The ADC, npADC ratio and the PRECISE score should be recorded for MRI-based AS.

Keywords: Prostate cancer; Active surveillance; Diffusion Magnetic Resonance Imaging; Molecular imaging.

Journal Pre-proof

Abbreviations and acronyms: PCa: prostate cancer AS: active surveillance mpMRI: multiparametric magnetic resonance imaging PI-RADS: Prostate Imaging Reporting And Data System DWI: diffusion-weighted imaging

ro of

ADC: apparent diffusion coefficient

PRECISE: Prostate Cancer Radiological Estimation of Change in Sequential Evaluation

-p

PSA: prostate specific antigen

re

T2-WI: T2-weighted imaging DCE: dynamic contrast-enhanced

lP

ROI: region of interest TZ: transitional zone

na

PZ: peripheral zone IQR: interquartile ranges

ur

ROC: receiver operating characteristic

Jo

ICC: intraclass correlation coefficient

Journal Pre-proof 1. Introduction Prostate cancer (PCa) is the most commonly diagnosed solid organ cancer in men in the United Kingdom, with around 40,000 cases diagnosed each year. [1] Patients with PCa confined to the prostate are classified into low, intermediate and high-risk categories and active surveillance (AS) is recommended for men with low or intermediate-risk disease. [1] The standard AS programmes include repeated blood tests, digital rectal

ro of

examination and standard biopsy.

There is compelling evidence that multiparametric magnetic resonance

-p

imaging (mpMRI) shows potential in identifying candidates for AS, who

re

may have little benefit from therapy, but still need to be monitored to allow prompt curative treatment if the disease shows signs of becoming

lP

harmful (radiological progression). [2–4]

Sanguedolce and colleagues explored which baseline mpMRI features

na

might be helpful for the refinement of AS inclusion criteria in 135 patients and concluded that a Prostate Imaging Reporting And Data System (PI-

ur

RADS) score >3 and an index lesion >10mm were strongly associated

Jo

with patient withdrawal from the AS programme. [5] Quantitative assessment of the change in diffusion-weighted imaging (DWI) over time is of interest as a way to detect who develop clinically significant disease whilst on surveillance. [4, 6–8] This technique measures the diffusion of water molecules within the extracellular space by the calculation of a quantitative parameter called apparent diffusion coefficient (ADC). [9] Because cancer cells are more tightly packed than benign cells, diffusion of water is restricted in the former and this corresponds to a higher signal on DWI and decreased signal on the ADC map (i.e. a lower ADC value than normal tissue). It has been reported that the ADC is a significant predictor of time to adverse histology on

Journal Pre-proof biopsy or radical treatment for PCa during AS. [10] However, the current reporting of studies of mpMRI during AS lacks rigour and standardisation. [2] In the UK, the National Institute for Health and Care Excellence guidelines recommend the use of serial mpMRI during AS, but they do not clearly report how often mpMRI should be performed and how serial mpMRI findings should be interpreted. [11] To address this need, the PRECISE (Prostate Cancer Radiological 2016. [12]

ro of

Estimation of Change in Sequential Evaluation) criteria were published in The aim of these recommendations, which are based on

international expert consensus, is to facilitate robust data collection from

-p

serial mpMRI during AS and to distinguish natural variability from changes

re

indicating true radiologic progression of PCa.

The PRECISE criteria recommend scoring the likelihood of radiological

lP

progression on follow up scans using a 1-to-5 scale (PRECISE score). A score of 1-2 indicates regression of a previously visible lesion, 3 denotes

na

stability and 4-5 corresponds to radiological progression. (Table 1)

ur

We conducted this pilot study in two academic centres (University College

Jo

London -UCL- and Sapienza), both highly experienced in prostate mpMRI reporting in order to address the following question: can DWI be used to differentiate those patients whose cancer progresses from those with stable disease? In order to do this, we firstly evaluated the inter-reader reproducibility of different ADC calculations from serial prostate MR scans and then explored if and how these are related to the PRECISE score.

Journal Pre-proof 2. Material and methods In this two-centre, retrospective study patient records and MR images were reviewed as part of an audit routinely performed for the internal evaluation of the AS service (Fig. 1). 2.1 Eligibility criteria and study design All prostate MR examinations were performed between January 2009 and

ro of

May 2019 and two radiologists (one from each centre; VP and FG with 11 and 7 years of experience in prostate MRI reporting, respectively) were involved in the study. Both radiologists had been actively involved in the

re

-p

discussion and drafting of the PRECISE recommendations in 2016. Inclusion criteria for this pilot study were: i) biopsy-proven PCa eligible

lP

for AS according to guidelines (i.e. Gleason 3+3 or 3+4) [11]; ii) prostate specific antigen (PSA) ≤15 ng/ml; iii) visible lesion on DWI at baseline

na

scoring ≥3 according to the PI-RADS version 2.1 guidelines [13] that was concordant with the histology using the six sectors scheme (i.e. base,

ur

midgland, and apex of the right and left prostate gland); iv) same MR

Jo

scanner (1.5T or 3T, same manufacturer) and protocol for baseline and follow-up scans.

Patients who had any treatment with any 5-alpha reductase inhibitors in the previous 12 months were excluded, as the use of such medications could reduce PCa conspicuity on DWI. [14] AS entry was defined as the initial diagnosis of PCa on biopsy and the follow-up period is updated to May 2019. To overcome any selection bias, eligible patients were randomly selected from each database. In this study, baseline MRI was not necessarily the first scan after entry into AS, as some of the patients had been scanned on different systems over the years. A specific inclusion criterion was to include only those patients who had been consecutively scanned on the

Journal Pre-proof same MR system (i.e. same magnet strength and same MR protocol including the same b values for each patient) both for baseline and followup scans (i.e. no other scans between them). At the beginning of the study, the two study radiologists identified the index lesion and assessed the PRECISE score for each patient in consensus. As per PRECISE guidelines, the lesion with the highest PIRADS score and volume (index tumour) was chosen for analysis, if multiple foci were detected in the same patient. Each PRECISE score was

ro of

the result of the comparison of visual evaluation (i.e. tumour conspicuity on different mpMRI sequences) and maximum lesion diameter on DWI for the peripheral zone (PZ) and on T2-WI for the transition zone (TZ) (in

-p

accordance with the strategy for lesion measurement provided in PI-RADS

re

V 2.1 guidelines) between baseline and follow-up scans. We applied the following specific interpretation to the PRECISE criteria, as

‘PRECISE 3’ (i.e. stability): either a scan with a stable lesion

na

i)

lP

per PRECISE recommendations:

‘PRECISE 4’ (i.e. progression): either a new focal lesion

Jo

ii)

ur

over time, or a persistent negative scan

(scored as PI-RADS ≥ 3) in a previous negative scan or a lesion with more suspicious MRI features (volume or conspicuity) since the last scan In case of only diffuse MRI changes in the prostate gland (as seen in prostatitis, for example), the MR scan was reported as ‘negative’ for the presence of focal lesions. In order not to introduce any bias, the ADC was not measured at this stage (i.e. the ADC did not influence the assessment of the PRECISE score). Each PRECISE score and lesion location were recorded on a separate spreadsheet by an independent observer. Then, at least 6 weeks later, each radiologist was provided with the list of cases (including the

Journal Pre-proof lesion location but not the PRECISE score) and the set of anonymised MR images of each patient in order to measure the ADC according to the region of interest (ROI) drawn by each radiologist independently. As per PRECISE recommendations, the radiologists were privy only to PSA, initial biopsy result and tumour location. 2.2 MR imaging technique Three different scanners were used: one 1.5T (Avanto, Siemens,

ro of

Erlangen, Germany) and one 3T systems (Achieva, Philips, Best, The Netherlands) at UCL and a 3T system (Discovery MR750, GE Healthcare, Milwaukee, USA) at Sapienza all using a multichannel surface phased-

-p

array body coil and no endorectal coils.

re

The protocols in both centres included T2-weighted (T2-WI), DWI (b values: 0, 100, 500, and 1000 s/mm2, and long b sequence: 1,400 s/mm2

lP

for 1.5T or 2,000 s/mm2 for 3T scanners) with ADC map calculation, and dynamic contrast enhanced (DCE) imaging, in accordance with

Jo

ur

2.3 Image analysis

na

international guidelines. [13, 15, 16]

All MR imaging data sets were reviewed using commercial image viewing software (Osirix X MD® v. 10.0.4; Geneva, Switzerland). Image quality was adequate to evaluate the ADC in all scans. All the available sequences (T2-WI, DWI and DCE) were used to accurately locate the lesion and median ADC values were obtained from ROIs traced on the ADC maps. The ROIs were positioned making reference to where the lesion was most conspicuous on the high b value sequence on DWI. The slice for the ROI had been previously identified in consensus during the initial session for the assessment of the PRECISE score.

Journal Pre-proof

In order to select an optimal reference for ADC normalisation, the two study radiologists copied and pasted another ROI of the same size in the non-cancerous PZ or TZ zone (according to tumour location) on the same slice in mirror position, and another ROI in the urine in the bladder lumen (Fig.2). These additional values were recorded and used to generate two parameters: the normalised prostatic ADC (npADC) and the normalised urinary ADC (nuADC) ratios, according to the formula: ADC (tumour)/ADC

ro of

(reference).

The concept of normalisation for ADC values is very important in order to minimise the variability between MR scanners and systems. As far this

-p

study is concerned, the npADC is more discriminating than the single ADC

re

value from the lesion, as one advantage of using the ADC from the noncancerous prostate as reference is that measurements are made easier

lP

with ROIs used for calculation placed on the same level of slice and because it is assumed that the adjacent tissue is subjected to the same

na

field heterogeneity and susceptibility effects than the lesion. Necrosis, blood vessels, and areas containing artefacts from bowel

Jo

ur

peristalsis were excluded from the ROIs. 2.4 Statistical analysis

Clinical and demographic data are reported using descriptive statistics. Continuous variables are expressed by median and interquartile ranges (IQR) and categorical data by frequencies and percentages. Interobserver consensus and agreement in measuring ADC was evaluated by Spearman’s correlation and intraclass correlation coefficients, and graphically inspected by Bland-Altman plots with confidence intervals at level 0.95. Measurements were averaged between the two observers for further analyses.

Journal Pre-proof Significant differences in ADC values and ratios between baseline and follow-up scans were verified using Kruskal-Wallis test statistics. Receiver operating characteristic (ROC) curves were generated and the areas under the curves were compared in order to differentiate between PRECISE 2 and 3 (i.e. radiological regression or stability) vs PRECISE 4 and 5 (i.e. radiological progression). To detect a difference of the median change of all parameters, between baseline and follow-up scans in the two groups we performed unpaired Ttest.

ro of

P values < 0.05 were considered to indicate a significant difference. All statistical analyses were performed using SPSS (version 25; SPSS,

Jo

ur

na

lP

re

-p

Chicago, Illinois, USA).

Journal Pre-proof 3. Results A total of 30 patients (fifteen scanned on 1.5T and fifteen on 3T MR systems) were included in this study. The median interval between baseline and follow-up mpMRI was 14 months (IQR: 12–18.75). There were 26/30 (87%) lesions in the PZ and 4/30 (13%) in the TZ. The median ROI size (averaged between the two readers) was 15 mm2 (IQR: 0.11–0.24) for baseline and 20 mm2 (IQR 0.14-0.41) for follow-up

ro of

mpMRI.

Table 2 shows the descriptive characteristics of all men included in the

-p

study. Overall, there were two (6%) PRECISE 2, fourteen (47%) PRECISE

re

3, eleven (37%) PRECISE 4 and three (10%) PRECISE 5 cases. At present 8/30 patients (27%) have received treatment. Of these, two

lP

(25%) have been treated with radical prostatectomy, five (63%) with

na

focal therapy and one (12%) with radiotherapy. The interobserver reproducibility between the two readers was very good

ur

both for baseline and follow-up ADC calculations, as shown by the high

Jo

Spearman’s rank and intraclass correlation coefficients (Table 3) and as graphically displayed in the Bland-Altman plots (Fig. 3). Given the high inter-reader reproducibility, the measurements were averaged between the two observers and used for the subsequent analyses. The overall median ADC values and normalised ADC ratios stratified by each PRECISE score at baseline and follow-up mpMRI are reported in Table 4. Splitting the overall population into radiological regression or stability (i.e. PRECISE 2 and 3) and radiological progression (i.e. PRECISE 4 and 5), significant differences were observed for lesion ADC and npADC ratio at follow-up mpMRI (p=0.025 and p=0.012, respectively), while

Journal Pre-proof there were no differences for all parameters at baseline and for the other values at follow-up mpMRI (Table 5). There were no differences of the change in all parameters between baseline and follow-up scans according to radiological regression or stability (i.e. PRECISE 2 and 3) vs radiological progression (i.e. PRECISE 4 and 5) (Table 6). The median lesion ADC and npADC ratio for each PRECISE score and

ro of

according to radiological regression or stability (i.e. PRECISE 2 and 3) vs radiological progression (i.e. PRECISE 4 and 5) at follow-up mpMRI are

-p

graphically displayed in Fig. 4.

re

According to ROC curves analysis (Fig. 5), a cut-off of 0.77 × 10−3 mm2/s for lesion ADC (AUC: 0.74; sensitivity: 68% and specificity: 64%) and a

lP

cut-off of 0.59 for npADC ratio (AUC: 0.770; sensitivity: 81% and

Jo

ur

PRECISE 4 and 5.

na

specificity: 71%) could differentiate between PRECISE 2 and 3 and

Journal Pre-proof 4. Discussion There is a need for reliable risk-assessment tools that can support the assessment of stability for patients on AS, and MRI-derived parameters such as the ADC represent an encouraging step in this direction. The main finding from our study is that the absolute ADC value of the lesion and the npADC ratio (normalised to non-cancerous prostatic tissue) on follow-up imaging are significantly different according to the presence PRECISE 2 and 3 vs PRECISE 4 and 5).

ro of

of radiological progression, defined as per PRECISE recommendations (i.e.

-p

First of all, there is evidence supporting the role of the ADC as a non-

re

invasive, quantitative tool for detection and follow-up of PCa, and a number of studies have shown that ADC values are inversely correlated

lP

with the Gleason score (i.e. PCa aggressiveness). [17, 18]

na

DWI is an essential component of mpMRI of the prostate and several previous studies have evaluated the utility of this technique also during

ur

AS. [4] Morgan and colleagues [19] investigated the change in tumour

Jo

volume over time in 151 patients on AS to determine whether baseline ADC and ADC changes were predictive of tumour growth. They concluded that change in T2-WI volume correlates with a change in ADC, and ADC may be used to identify patients with clinically significant growth, suggesting a 5.8% reduction in ADC as a possible threshold (specificity: 77%; sensitivity:54.9%) for indicating volume progression. One of the biggest limitations in DWI is the inter-observer variability when assessing the ADC, and reader experience is very important at this regard. We found in our study that two expert radiologists had a very good agreement in calculating the ADC values both at baseline and follow-up mpMRI.

Journal Pre-proof At present, there are no clear recommendations on how (e.g. drawing a single region of interest or by planimetry) and which (absolute vs normalised) ADC values should be calculated. It is known that absolute ADC values vary from one MR system to the other and that they are dependent on the number of b values acquired. There is also considerable inter- and intra-patient variability. [20] We aimed to minimise these drawbacks by including only those patients who had been scanned on the same MR system (i.e. same magnet strength and same MR protocol including the same b values) at both time

ro of

points. We also used the ratio of tumour ADC to that of normal prostatic tissue for overcoming the variability in absolute ADC values (i.e. we assume that the variability of ADC in normal prostatic tissue is equivalent

re

-p

to that within tumours).

Different authors have investigated the potential role of the ADC ratio in

lP

the evaluation of PCa using different reference standards for ADC normalisation, including normal prostatic tissue [17] and urine in the

na

bladder.[21]

We found a significant difference for radiological progression (defined as

ur

PRECISE 2 and 3 vs PRECISE 4 and 5) for tumour ADC (0.91 vs 0.73 x

Jo

10-3 mm2/s; p=0.025) and the npADC ratio (0.68 vs 0.53; p=0.012) at follow-up mpMRI, but no significant results were observed at baseline mpMRI. No difference was seen at baseline and follow-up mpMRI when tumour ADC was normalised to urine in the bladder (nuADC), even though a trend towards statistical significance was observed at follow-up mpMRI (0.52 vs 0.38, p=0.007). This means that we cannot yet infer that baseline ADC values can predict the likelihood of radiological progression, and further research is needed at this regard. In addition to the aforementioned results, we should recall that both biopsy and PSA measurements are prone to variability (sampling errors

Journal Pre-proof and natural PSA fluctuation), and there is need of novel markers to detect PCa progression on AS, including image-derived parameters. To date, no standard ADC cut-offs have been established to assess the radiological progression, but our study provides a first answer to this, as we found two optimal cut-offs to discriminate between radiological stability and radiological progression (ADC: 0.77 × 10−3 mm2/s and npADC ratio 0.59). Henderson and colleagues [10] reported that ADC is a significant predictor of AS disqualification (i.e. time to adverse histology on biopsy or radical treatment) over nine years of follow up, as patients

ro of

with a baseline ADC < 972 mm2/s in their AS cohort progressed, on average, seven years earlier. However, differently from our study, the ADC was calculated only on baseline MRI and they did not perform any

re

-p

kind of ADC normalisation.

Our findings hint that the ADC from follow-up DWI could assist in the

lP

identification of patients with radiological progression (who should be biopsied or offered treatment) and, at the same time, could help in

na

identifying those without radiological progression (who could benefit from

ur

a more conservative approach, i.e. clinical and mpMRI follow up).

Jo

A number of methodological limitations apply to our study. First is the relatively small cohort of patients, but we aimed to conduct a pilot study on a specific cohort where we minimised the variability related to magnet strengths, vendors and protocols. Fifteen patients (50%) were exclusively scanned on a 1.5T and the other fifteen (50%) exclusively on a 3T scanner, using always the same machine and the same protocol for each patient. In order to achieve good quality results and reliable ADC values, it is fundamental that the MR systems are regularly checked, and the MR protocols adhere strictly to the PI-RADS v. 2.1 guidelines. Therefore, good communication between the radiologist and the radiographer is essential, and if there are artefacts that can compromise DWI quality (e.g. rectal air), measures should be taken to rectify the

Journal Pre-proof problem and the sequence should be repeated, as recommended in the PI-RADS v. 2.1 guidelines. The second limitation is the absence of tissue verification by means of radical prostatectomy, but this is common in AS cohorts. Only 12/30 (40%) patients in our population underwent rebiopsy (eight of whom with a targeted approach), but resampling was often triggered by apparent tumour growth on mpMRI or by adverse PSA kinetics without MRI changes. However, this study was specifically designed to investigate the correlation of ADC with radiological progression, which was exclusively

ro of

based on MR features and not on histological progression. There were only four lesions in the TZ (that is more heterogeneous than the PZ from a histological point of view) and it is possible that different

-p

evaluations would have arisen if the number of TZ lesions had been

re

higher.

Moreover, due to the inclusion criteria requiring visible lesions on DWI at

lP

baseline and follow up, there were no scans scored as PRECISE 1. This is why we created two groups (PRECISE 2 and 3 vs PRECISE 4 and 5) that

na

were fairly balanced in terms of population (16 vs 14), without altering the clinical message behind (i.e. patients with radiological stability could

ur

be followed up without the need of resampling, differently from those

Jo

showing radiological progression). Lastly, it should be noted that changes in ADC and an increase in subjective tumour conspicuity (that is one of the main drivers for defining radiological progression in the PRECISE score) are closely correlated, as the ADC map is the main way of assessing most tumours (especially in the PZ). In conclusion, our initial results hint that the ADC correlates with radiological progression, in line with previous studies [14, 22]. In addition to this, the calculation of the npADC ratio using our method (Fig. 6) could be easily performed during daily clinical practice and used as an additional image-derived quantitative parameter that can assist the Radiologist in

Journal Pre-proof the assessment of radiological progression by means of PRECISE score. The ADC, npADC ratio and the PRECISE score should be recorded for MRI-

Jo

ur

na

lP

re

-p

ro of

based AS.

Journal Pre-proof Acknowledgements We thank all the patients who greatly contributed to the realisation of this study. This study has received funding by the European School of Radiology (ESOR) under the umbrella of the 2019 BRACCO fellowship and the ERASMUS+ programme (University College London) Dr. Francesco Giganti is funded by the UCL Graduate Research

ro of



Scholarship, the Brahm PhD scholarship in memory of Chris Adams, the ESOR 2019 BRACCO fellowship and the ERASMUS + programme (UCL). Prof. Shonit Punwani receives research support from the United

-p



Biomedical Research Centre.

Dr. Alex Kirkham receives research support from the United

lP



re

Kingdom’s National Institute of Health Research (NIHR) UCLH/UCL

Kingdom’s National Institute of Health Research (NIHR) UCLH/UCL •

na

Biomedical Research Centre.

Prof. Mark Emberton receives research support from the United

ur

Kingdom’s National Institute of Health Research (NIHR) UCLH/UCL

Jo

Biomedical Research Centre. He is an NIHR Senior Investigator.

Journal Pre-proof References 1.

Tosoian JJ, Carter HB, Lepor A, Loeb S. Active surveillance for prostate cancer: Current evidence and contemporary state of practice. Nat Rev Urol 13:4 (2016) 205–215. doi: 10.1038/nrurol.2016.45

2.

Schoots IG, Petrides N, Giganti F, et al. Magnetic resonance imaging in active surveillance of prostate cancer: A systematic review. Eur Urol 67:4 (2015) 627–636. doi: 10.1016/j.eururo.2014.10.050 Stavrinides V, Giganti F, Emberton M, Moore CM. MRI in active

ro of

3.

surveillance: a critical review. Prostate Cancer Prostatic Dis 22:1 (2019) 5-15. doi: 10.1038/s41391-018-0077-2

Giganti F, Moore CM. Magnetic resonance imaging in active

-p

4.

surveillance—a modern approach. Transl Androl Urol 7:1 (2018) Sanguedolce F, Petralia G, Sokhi H, et al. Baseline Multiparametric

lP

5.

re

116–131. doi: 10.21037/tau.2017.12.23. MRI for Selection of Prostate Cancer Patients Suitable for Active

na

Surveillance: Which Features Matter? Clin Genitourin Cancer 16:2 (2018) 155-163.e6. doi: 10.1016/j.clgc.2017.10.020 Malayeri, AA, El Khouli RH, Zaheer A, et al. Principles and

ur

6.

Applications of Diffusion-weighted Imaging in Cancer Detection,

Jo

Staging and Treatment Follow-Up. Radiographics 31:6 (2011) 1773– 1791. doi: 10.1148/rg.316115515 7.

Algohary A, Viswanath S, Shiradkar R, et al. Radiomic features on MRI enable risk categorization of prostate cancer patients on active surveillance: preliminary findings. J Magn Reson Imaging (2018) (in press) doi: 10.1002/jmri.25983.

8.

Maurer MH, Heverhagen JT. Diffusion weighted imaging of the prostate—principles, application, and advances. Transl Androl Urol 6:3 (2017) 490–498. doi: 10.21037/tau.2017.05.06.

9.

Tan CH, Wang J, Kundra V.Diffusion weighted imaging in prostate cancer. Eur Radiol 21:3 (2011) 593–603. doi: 10.1007/s00330-010-

Journal Pre-proof 1960-y. 10.

Henderson DR, De Souza NM, Thomas K, et al. Nine-year Follow-up for a Study of Diffusion-weighted Magnetic Resonance Imaging in a Prospective Prostate Cancer Active Surveillance Cohort. Eur Urol 69:6 (2016) 1028–1033. doi: 10.1016/j.eururo.2015.10.010.

11.

National Institute for Health and Care Excellence (2019) Prostate cancer: diagnosis and management (NICE Guideline 131). Available at:https://www.nice.org.uk/guidance/ng131/chapter/recommendati ons#active-surveillance [Accessed 02/09/2019]. Moore CM, Giganti F, Albertsen P, et al. Reporting Magnetic

ro of

12.

Resonance Imaging in Men on Active Surveillance for Prostate Cancer: The PRECISE Recommendations - A Report of a European

13.

re

10.1016/j.eururo.2016.06.011.

-p

School of Oncology Task Force. Eur Urol 71:4 (2017) 648–655. doi: Turkbey B, Ronsenkrantz AB, Haider M, et al. Prostate Imaging

lP

Reporting and Data System Version 2.1: 2019 Update of Prostate Imaging Reporting and Data System Version 2. Eur Urol 76:3 (2019) 14.

na

340-351 doi:10.1016/j.eururo.2019.02.033 Giganti F, Moore CM, Robertson NL, et al. (2017) MRI findings in

ur

men on active surveillance for prostate cancer: does dutasteride

Jo

make MRI visible lesions less conspicuous? Results from a placebocontrolled, randomised clinical trial. Eur Radiol 27:4767–4774. 15.

Barentsz JO, Richenberg J, Clements R, et al. ESUR prostate MR guidelines 2012. Eur Radiol 22:4 (2012) 746–757. doi: 10.1007/s00330-011-2377-y.

16.

Weinreb JC, Barentsz JO, Choyke PL, et al. PI-RADS Prostate Imaging - Reporting and Data System: 2015, Version 2. Eur Urol 69:1 (2016) 16–40. doi: 10.1016/j.eururo.2015.08.052.

17.

De Cobelli F, Ravelli S, Esposito A, et al. Apparent diffusion coefficient value and ratio as noninvasive potential biomarkers to predict prostate cancer grading: Comparison with prostate biopsy and radical prostatectomy specimen. AJR Am J Roentgenol 204:3

Journal Pre-proof (2015) 550–557. doi: 10.2214/AJR.14.13146. 18.

deSouza NM, Riches SF, VanAs NJ, et al. Diffusion-weighted magnetic resonance imaging: a potential non-invasive marker of tumour aggressiveness in localized prostate cancer. Clin Radiol 63:7 (2008) 774–782. doi: 10.1016/j.crad.2008.02.001

19.

Morgan VA, Parker C, MacDonald A, et al.Monitoring tumor volume in patients with prostate cancer undergoing active surveillance: Is MRI apparent diffusion coefficient indicative of tumor growth? AJR Am J Roentgenol 209:3 (2017) 620–628. doi:

20.

ro of

10.2214/AJR.17.17790.

Barrett T, Priest AN, Lawrence EM, et al. Ratio of tumor to normal prostate tissure apparent diffusion coefficient as a method for

-p

quantifying DWI of the prostate. AJR Am J Roentgenol 205:6 (2015) 21.

re

W585-93. doi: 10.2214/AJR.15.14338.

Rosenkrantz AB, Kopec M, Kong X, et al. Prostate Cancer vs. Post-

lP

Biopsy Hemorrhage: Diagnosis with T2- and Diffusion-Weighted Imaging. J Magn Reson Imaging 31:6 (2010) 1387–1394. doi: Morgan VA, Riches SF, Thomas K, et al. Diffusion-weighted magnetic

ur

resonance imaging for monitoring prostate cancer progression in patients managed by active surveillance. Br J Radiol 84:997 (2011)

Jo

22.

na

10.1002/jmri.22172.

31-7. doi: 10.1259/bjr/14556365.

Journal Pre-proof Table 1: Assessment of likelihood of radiological progression on magnetic resonance imaging in men on active surveillance for prostate cancer (PRECISE score) Table 2 - Descriptive characteristics for each man included in the study (n=30) Table 3 – Spearman’s rank correlation (ρ) and intraclass correlation

ro of

coefficients at baseline and follow-up mpMRI Table 4 - ADC values (x10-3 mm2/s) and normalised ADC ratios stratified

-p

by PRECISE score at baseline and follow-up mpMRI

re

Table 5 - ADC values (x10-3 mm2/s) and normalised ADC ratios stratified by grouped PRECISE score (i.e. radiological regression/stability -PRECISE

lP

2 and 3- vs radiological progression -PRECISE 4 and 5-) at baseline and

na

follow-up mpMRI

Table 6 – Differences of the median change (Δ)of all parameters between

ur

baseline and follow-up scans stratified by grouped PRECISE score (i.e.

Jo

radiological regression/stability vs radiological progression).

Journal Pre-proof Figure 1 - Diagram indicating multiparametric MRI schedule undertaken based on baseline MRI status. The timing of MRI on active surveillance was based on both baseline risk and changes during follow up. Figure 2 – Multiparametric magnetic resonance imaging of a 77-year-old patient with biopsy-proven prostate cancer in the left midgland peripheral zone. T2-weighted (A), dynamic-contrast enhanced (B) and diffusion weighted (C) imaging confirm the presence of the lesion (arrows). Three different regions of interest (of the same size and area) from the

ro of

apparent diffusion coefficient map were drawn on the lesion and normal prostatic tissue (D) and on the urine in the bladder (E).

-p

Figure 3 - Bland-Altman plots representing the interobserver

re

reproducibility between the two readers for the different ADC and normalised ADC ratio values, both at baseline (A, B, C) and at follow-up

lP

magnetic resonance imaging (D, E, F). The centre line represents the mean of differences, the top line shows the upper 95% limit of

na

agreement, and the bottom line shows the lower 95% limit of agreement, with the mean difference between the long- and short-axis measurements

Jo

ur

(±1.96 times the standard deviation). Figure 4 - Boxplots showing lesion ADC (A, B) and normalised prostatic ADC (npADC) values (C, D) at follow-up magnetic resonance imaging as function of each single PRECISE score (A, C) and according to radiological regression/stability (PRECISE 2 and 3) or radiological progression (PRECISE 4 and 5) (B, D). Figure 5 - ROC curves for the detection of radiological progression on the basis of lesion ADC (blue, long-dashed line) and normalised prostatic ADC (npADC) (red, short-dashed line) values at follow-up magnetic resonance imaging.

Journal Pre-proof Figure 6 – Multiparametric magnetic resonance imaging of a lesion in the peripheral zone on T2-weighted imaging (A, arrow) and ADC map (B-C) and in the transitional zone on T2-weighted imaging (D, arrow) and ADC map (E-F) from two different patients on active surveillance for prostate cancer. The image includes the median ADC values (lesion -blue circles- and non-cancerous tissue -yellow and green circles-) obtained from the different regions of interest, and the corresponding

Jo

ur

na

lP

re

-p

ro of

npADC ratios for the lesion in the peripheral (C) and transitional (D) zone.

Journal Pre-proof

Author contribution

Francesco Giganti: Conceptualization, Methodology, Formal analysis, Investigation, Data curation, Writing – Original draft, Funding acquisition Martina Pecoraro: Conceptualization, Investigation, Data curation, Writing – Review and Editing

Riccardo Campa: Data curation Francesco Del Giudice: Data curation

ro of

Davide Fierro: Software

Shonit Punwani: Writing – Review and Editing Clare Allen: Writing – Review and Editing

-p

Alex Kirkham: Writing – Review and Editing

re

Mark Emberton: Supervision, Funding acquisition, Writing – Review and Editing Carlo Catalano: Supervision, Funding acquisition, Writing – Review and Editing

lP

Caroline M Moore: Supervision, Funding acquisition, Writing – Review and Editing,

Jo

ur

Review and Editing

na

Valeria Panebianco: Conceptualization, Supervision, Funding acquisition, Writing –

Journal Pre-proof Table 1: Assessment of likelihood of radiological progression on magnetic resonance imaging in men on active surveillance for prostate cancer (PRECISE score) PRECISE

Assessment of likelihood of radiological progression

score Resolution of previous features suspicious on MRI

2

Reduction in volume and/or conspicuity of features suspicious for prostate cancer

3

Stable MRI appearance: no new focal/diffuse lesions

4

Significant increase in size and/or conspicuity of features suspicious for prostate cancer

5

Definite radiologic stage progression (ECE, SV involvement, LN involvement, metastasis)

ro of

1

Legend: MRI=Magnetic Resonance Imaging; ECE=extracapsular extension;

Jo

ur

na

lP

re

-p

SV=seminal vesicle; LN=lymph node)

Journal Pre-proof Table 2 – Descriptive characteristics for each man included in the study (n=30) Age (year s)

MR syste m

Baseli ne PSA (ng/ml )

Baseli ne PSA densit y

Gleas on score at entry

Type of biop sy at entry

Follo w-up PSA (ng/m l)

Follo w-up PSA densi ty

Gleas on score at rebiopsy

Type of biop sy at rebiop sy

1

66

3T

5.2

0.11

3+3

S

3.64

0.07

3+3

S

2

69

3T

8.09

0.11

3+3

T

4.57

0.06

-

-

3+3

T

3+3

T

3+4

S

-

-

1.6

0.03

-

-

0.5

0.01

-

-

-

-

0.17

-

-

0.05

-

55 61

3T 3T

6.07 5.58

0.12 0.08

5

75

3T

1.95

0.07

3+3

S

6

59

3T

0.5

0.01

3+3

S

3+3

T S

8 9 10 11 12

70

3T

9.22

0.22

62

3T

5.5

0.14

3+4

61

3T

2.93

0.05

3+3

8.8

0.07

3+3

60

3T

7.9

0.17

3+3

S

3+3

T T

76

3T

5.63

0.13

73

3T

3.23

0.16

3+3

14

66

3T

1.8

0.06

3+3

8.01 3 11.96

S T

64

3T

9.4

0.23

71

1.5T

6.8

0.07

17

61

1.5T

9.4

0.14

lP

15

12.15

S

3T

13

5.18

S

64

3+3

18

67

1.5T

11.3

0.12

75

1.5T

11.86

20

67

1.5T

4.87

22

52 55

1.5T 1.5T

23

63

1.5T

24

73

1.5T

25

7

0.25 0.23

Jo

21

0.16

3.3

0.07

0.2

0.09

8.1

9.45

0.17 0.19

-

3+4

T

-

-

2.72

0.15

-

-

0.73

0.02

-

-

0.04

-

S

11.29

0.13

3+4

S

3+3

T

12.8

0.14

3+3

T

3+3

S

11.4

0.15

3+3

S

3+3

S

7.7

0.4

-

-

3+3

S

-

-

3+3

S

-

-

6.15

0.19

3+4

S+T

9.4

0.29

3+4

S

6.4

0.2

3+4

S

5.5

0.19

3+3

S

77

1.5T

7

0.14

3+3

26

62

1.5T

7.5

0.15

3+3

27

65

1.5T

6

0.08

S

12.41 5.2

0.34 0.09

13 12 14 13 22 22 21 16 14 18 11 25

T

5.5

S

16

-

3+4

3+3

13

-

0.21

ur

19

0.06

10

na

16

3+4

0.18

re

7

8.89

ro of

4

-p

3

Time interva l betwe en MR scans (month s)

12 12 12 23 21 13 16 23 13 12

-

10.25

0.2

-

S

9

0.18

3+4

T

3+3

S

11

0.14

3+4

T

18

28

50

1.5T

5.4

0.13

3+3

S

5.86

0.14

-

-

29 30

58

1.5T

5.7

0.2

3+4

3.38

0.13

-

81

1.5T

10.6

0.11

3+4

S S

13

0.12

4+3

T

Medi an (IQR)

65

-

6.03 (5.28.09)

0.13 (0.080.17)

-

-

8.05 (4.57 -11)

0.14 (0.07 0.19)

-

-

12 18 12 10 18 14 (1218.75)

Time on AS * (year s)

PRECI SE score

4.09

2

2.58

2

3.08

3

4.41

3

1.58

3

4

3

4.92

3

8.19

3

12.6 4 10.5 8 7.67

3 3 3

2

4

3.58

4

2.75

4

12.7 6 11.6 1 11

4 3 3

4.66

3

5.97

3

7.73

3

9.21

4

7.96

4

4.93

4

4.06

4

11.7 9 7.92

4 4

6.03

4

1.11

5

1.03 4.52

5

4.9 (3.45 8.44)

-

5

Legend – MR: magnetic resonance; PSA: prostate specific antigen; IQR: interquartile ranges. * AS entry was defined as the date of the first positive biopsy

Journal Pre-proof Table 3 – Spearman’s rank correlation coefficients (ρ) and intraclass correlation coefficient at baseline and follow-up mpMRI. CI

ICC

CI

0.745-0.937

0.926

0.833-0.966

0.925

0.847-0.964

0.962

0.907-0.983

0.887

0.775-0.945

0.976

0.950-0.988

0.830

0.671-0.916

0.932

0.857-0.967

0.914

0.826-0.967

0.973

0.943-0.987

0.963

0.923-0.982

0.983

0.964-0.992

ro of

Baseline normal ADC Baseline lesion ADC Baseline urine ADC Follow-up normal ADC Follow-up lesion ADC Follow-up urine ADC

Spearman’s rho 0.871

-p

Legend - CI: confidence intervals; ICC: intraclass correlation coefficient; ADC: apparent diffusion coefficient

re

Table 4 – ADC values (x10-3 mm2/s) and normalised ADC ratios stratified by PRECISE score at baseline and follow-up mpMRI

lP

PRECISE 2 (n= 2)

npADC

Followup mpMRI

Jo

nuADC

ur

Bladder ADC

Normal prostatic tissue ADC Lesion ADC npADC Bladder ADC nuADC

1.36 (1.051.67) 0.73 (0.590.87) 0.54 (0.520.56) 1.63 (1.511.76) 0.44 (0.390.50) 1.39 (1.271.52) 0.97 (0.741.21) 0.69 (0.580.80) 1.74 (1.422.07) 0.55 (0.520.59)

na

Normal prostatic Baseline tissue ADC mpMRI Lesion ADC

PRECISE 3 (n= 14)

PRECISE 4 (n= 11)

PRECISE 5 (n= 3)

1.43 (1.241.56) 0.94 (0.861.02) 0.66 (0.600.74) 1.73 (1.611.77) 0.55 (0.500.58) 1.30 (1.151.38) 0.91 (0.760.97) 0.68 (0.600.78) 1.73 (1.661.81) 0.51 (0.440.58)

1.49 (1.291.60) 0.82 (0.740.97) 0.61 (0.510.69) 1.75 (1.611.22) 0.46 (0.350.60) 1.37 (1.101.52) 0.70 (0.600.80) 0.53 (0.450.59) 1.51 (1.222.12) 0.36 (0.330.53)

1.26 (1.131.63) 0.85 (0.631.02) 0.63 (0.500.75) 1.68 (1.331.75) 0.51 (0.470.58) 1.38 (1.161.39) 0.80 (0.580.87) 0.58 (0.420.75) 1.78 (1.771.82) 0.45 (0.320.49)

Legend – Data are medians with interquartile ranges in parentheses. MpMRI: multiparametric magnetic resonance imaging; ADC: apparent diffusion coefficient; npADC: normalised prostatic ADC; nuADC: normalised urinary ADC

Journal Pre-proof

Table 5 – ADC values (x10-3 mm2/s) and normalised ADC ratios stratified by grouped PRECISE score (i.e. radiological regression/stability vs radiological progression) at baseline and follow-up mpMRI p

Normal prostatic tissue ADC Lesion ADC

1.43 (1.23-1.58)

1.47 (1.25-1.60)

0.662

0.90 (0.83-1.01)

0.83 (0.75-0.95)

0.163

npADC

0.64 (0.60-0.73)

0.62 (0.54-0.65)

0.197

Bladder ADC nuADC Normal prostatic tissue ADC Lesion ADC npADC

1.73 (1.61-1.78) 0.54 (0.49-0.60) 1.31 (1.21-1.38)

1.71 (1.58-1.94) 0.48 (0.41-0.48) 1.37 (1.13-1.44)

0.803 0.151 0.724

0.91 (0.77-0.99) 0.68 (0.60-0.76)

0.73 (0.59-0.83) 0.53 (0.48-0.63)

0.025 0.012

Bladder ADC nuADC

1.74 (1.57-2) 0.52 (0.44-0.57)

1.73 (1.41-1.90) 0.38 (0.34-0.55)

0.560 0.070

ro of

Follow-up mpMRI

PRECISE 4 and 5 (n= 14)

-p

Baseline mpMRI

PRECISE 2 and 3 (n= 16)

lP

re

Legend – Data are medians with interquartile ranges in parentheses. MpMRI: multiparametric magnetic resonance imaging; ADC: apparent diffusion coefficient; npADC: normalised prostatic ADC; nuADC: normalised urinary ADC

PRECISE 4 and 5 (n= 14)

p

-7.43 (-16.02 — -0.33) -5.39 (-13.87 — 5.78)

-5.57 (-19.70 — 8.39) -14.79 (-21.02 — -3.22)

0.81 0.08

1.79 (-5.17 — 11.07) -1.53 (-7.15 — 14.42) -6.52 (-18.97 — 5.93)

-11.12 (-19.38 — 3.49) -2.71 (-14.43 — 2.79) -9.95 (-22.79 — -1.05)

0.14 0.22 0.98

ur

Δ npADC Δ bladder ADC Δ nuADC

PRECISE 2 and 3 (n= 16)

Jo

Δ prostatic tissue ADC Δ lesion ADC

na

Table 6 – Differences of the median change of all parameters between baseline and follow-up scans stratified by grouped PRECISE score (i.e. radiological regression/stability vs radiological progression).

Legend – Data are medians with interquartile ranges in parentheses. ADC: apparent diffusion coefficient; npADC: normalised prostatic ADC; nuADC: normalised urinary ADC

Journal Pre-proof

Highlights 

ADC values are inversely correlated with the PRECISE score.



Lower ADC values on follow-up imaging are associated with radiological progression. The role of ADC in serial MR scans during AS should be further

ur

na

lP

re

-p

ro of

investigated.

Jo



Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6