Use of diffusion weighted imaging (DWI) MRI in localizing prostate carcinoma

Abstract

Objective: The aim of this study was to evaluate the role of diffusion-weighted imaging (DWI) and Apparent Diffusion Coefficient (ADC) MRI techniques in the detection and localization of prostate carcinoma in Pusat Perubatan Universiti Kebangsaan Malaysia (PPUKM). Materials and Methods: DWI ( $b$ value $=$ 800 s/mm ${}^{2})$ and ADC at 3T were evaluated retrospectively in 78 male patients with suspicion of prostate cancer. The prostate MRI was evaluated after dividing each prostate gland into 6 sections ( $n=$ 462). Six sections were excluded as biopsy result came back as unsatisfactory. Minimum ADC values of each section in the peripheral zone were measured. The MRI finding was compared with biopsy result obtained from each section. Results: The mean minimum ADC values of malignant sextants were significantly lower than that of noncancerous tissue, with a significant negative correlation between the ADC value and the Gleason score. The sensitivity and specificity for DWI $+$ ADC were 52.14% and 93.48% respectively. There is significant difference between malignant and benign ADC value ( $p=$ 0.0000). The area under the receiver operating characteristic curve for the detection and localization of prostate cancer within the peripheral zone is 0.788 for ADC, which represents a fair to good test. The cut off ADC value between benign and malignant tissue is 1174.5 mm ${}^{2}$ /s at 93.48% specificity. Conclusion: The current DWI and ADC technique used in our center is very specific and is valuable in detecting, localizing, grading prostate carcinoma and guiding targeted biopsy, with areas of lowest ADC values can indicate the regions to be biopsied.

Keywords

DWI MRI ADC prostate carcinoma

1. Introduction

Diffusion weighted imaging (DWI) sequence has been shown to give better detection of tumour in prostate gland. In the last decade, DWI technique has been part of multiparametric MRI (mp-MRI) protocol for prostate imaging, together with dynamic contrast-enhanced (DCE), spectroscopy, T1- and T2-weighted (T1W and T2W) imaging [1]. DWI imaging method allows mapping of the diffusion process of water molecules in biological tissue. In cancerous tissue, water molecules are restricted within the cells bounded by the cell membrane whereas in normal tissue, water molecules are free to move through the cell membrane to extracellular space in free diffusion. As DWI sequence detects the water molecule as bright signal, cancerous tissue with restricted water molecule movement will appear brighter on DWI compared to free water molecules in the background tissue. Visual assessment of relative tissue signal intensity at DWI can be applied for tumour detection, characterization and evaluation of treatment response in prostate carcinoma [2].

Apart from visual assessment, DWI images can also be assessed quantitatively using calculation of ADC value. Areas of restricted diffusion in highly cellular areas show low ADC values compared with less cellular areas that return higher ADC value. On MRI image, areas of restricted diffusion (i.e. tumour tissue) will appear bright on increasing $b$ value in DWI image and dark on the ADC map [3].

Our study address how accurate DWI/ADC technique is in detection of prostate carcinoma, illustrate its use in guiding targeted biopsy and what is the cut off ADC value between benign and malignant tissue.

2. Methodology and technique

2.1 Study type

We conducted a retrospective study in the Department of Diagnostic Imaging of Universiti Kebangsaan Malaysia Medical Center. The study was approved by the hospital’s ethical committee, with no ethical issue to declare. MRI prostate images that had prostate biopsy done from 2013 until 2014 were collected and reviewed. There were 78 patients who met the inclusion criteria and were enrolled into this study.

2.2 MRI technique

MRI prostate was done using a 3 T Siemens machine (A Tim System MAGNETOM Verio; SIEMENS, Germany). Patient fasted 6–8 hours before examination. No enema or bowel preparation was required. The maximum gradient amplitude of 33 mT/m and maximum slew rate of 120 mT/m/sec, were used using a pelvic 130 phased-array coil. Multiparametric MRI prostate sequences which include T1W, T2W, DWI, ADC, dynamic contrast enhanced and spectroscopy were performed.

The imaging parameters for the fast spin-echo T2-weighted images were TR/TE, 4680/76 ms; echo-train length,19; bandwidth, 31.25 kHz; field of view, 14 cm; slice thickness, 3 mm; gap, 0.5 mm; number of excitations, 3; phase-encoding direction, right to left; and matrix 384 $\times$ 384. Axial echoplanar DWI was performed using slice locations similar to those used for the T2-weighted sequence using the following parameters: b value, 0 s/mm ${}^{2}$ , 100 s/mm ${}^{2}$ and 800 s/mm ${}^{2}$ ; TR/TE, 8200/93; bandwidth, 250 kHz; field of view, 24 cm; slice thickness, 4 mm; number of excitations, 8; matrix, 178 $\times$ 1252.

Data regarding DWI signal appearance, ADC value, prostate specific antigen (PSA) value and Gleason biopsy score of the patients were recorded. The result is correlated with histopathological finding for section based prostate biopsy.

2.3 Data interpretation

All images were reviewed by a single radiologist with 4 years of dedicated experience in interpreting prostate MRI, using Osirix. The peripheral zone of the prostate was divided into six regions: base, middle, and apex and left and right halves. The base was defined as the upper third of the prostate, which extended from the vesical margin of the prostate; the mid region was defined as the central third; the apex was defined as the remaining inferior third. For all patients, T2W and DWI sequence was reviewed first. On the T2WI, sections with areas of any nodular low signal intensity in the peripheral zone were assigned as malignant, whereas mild low signals with linear or feathered appearances were accepted as benign. For DWI alone, the ADC maps were constructed on the workstation. The ADC values of each region in the peripheral zone were measured by placement of the region of interest (ROI) circle on hypointense areas of the ADC map images. The lowest ADC values of at least three ROI measurements were recorded for each Section 2-weighted images were reviewed in conjunction with the DWI and ADC maps, without knowledge of the biopsy results. For T2WI $+$ DWI, any low T2 intensity with abnormal ADC values or a mass-like appearance on T2WI with borderline ADC values was assigned as malignant, whereas mild low signals in a linear or feathered appearance with borderline ADC values were accepted as benign. For ADC evaluation, lesion with low intensity and low ADC value compared to surrounding tissue is taken as malignant. Lesion detected on DWI and ADC is compared to biopsy result and PSA value.

2.4 Statistical analysis

Statistical analysis were performed using Statistical Package for Social Science (SPSS) for Mac (IBM SPSS Statistics 22). The ADC values results are expressed as the mean $\pm$ standard deviation. Student’s $t$ test was used to determine significant differences in the mean ADC value between benign and malignant sextants in the peripheral gland. For each patient with prostate cancer, the mean of the minimum ADC values of malignant sections was calculated. $P$ value of $<$ 0.05 was considered to indicate statistically significant differences. Relationships between mean ADC values in cancerous sections and total Gleason scores were assessed using Spearman’s correlation test. For DWI, the cut-off ADC value distinguishing between benign and malignant lesions were defined using receiver operating characteristic (ROC) curve analysis. Sensitivities, specificities, positive predictive values, negative predictive values, and accuracies of detecting prostate cancer were calculated for each MRI protocol.

3. Result

There were 78 patients enrolled into this study. The patients were within the age range of 55 to 111 years old (mean age of 68.4 years old). These patients have PSA level between 0.49 ng/ml to 156.56 ng/ml (mean of 17.23 ng/ml) with Gleson score of 4 to 9 (mean 7.06). Majority of the patients were diagnosed with adenocarcinoma (53%). The rest of the patients were diagnosed with BPH (45%), BPH with prostitis (1%) and high grade intraepithelial neoplasia (1%). From the 78 patients, a total of 468 prostate sections were obtained and analyzed, but six prostate sections histopathological samples were unsatisfactory and thus excluded leaving a total of 462 sections 94 prostate sections were detected as cancer positive by MRI and the remaining 364 prostate sections were cancer negative by MRI. Out of 94 MRI cancer positive Sections 73 sections were biopsy proven to be malignancy (TP) and the rest of 21 sections were benign (FP). Three hundred and sixty-four prostate sections were detected negative for cancer on MRI. Out of these, 67 sections were biopsy proven positive for cancer (FN) and the remaining 301 sections were proven negative for cancer (TN) (Table 2). From these results, DWI/ADC MRI shows 52.14% sensitivity and 93.48% specificity in detecting prostate cancer lesion, with positive predictive value (PPV) of 77.66% and negative predictive value (NPV) of 82.69%. The accuracy of DWI/ADC MRI in detecting prostate carcinoma is 80.95% (Table 1).

4. Discussion

Prostate cancer accounts for 25% of all male cancers and that the prevalence of prostate cancer increases with age. Our demographic data shows that majority of patients with prostate cancer aged between 50–80 years old, with mean age of 68.4 years old. Conventionally, detection of prostate carcinoma is done using the plasma prostate specific antigen (PSA), digital rectal examination (DRE) or transrectal ultrasound-guided biopsy. Although PSA demonstrate high sensitivity, its specificity is low, owing to false positive PSA elevation with non-cancerous conditions such as benign prostatic hyperplasia (BPH) and chronic prostatitis [4, 5]. As in our study, most of the patients had a low to slightly higher PSA value when diagnosed with prostate carcinoma. Previous study shows that MRI is a more accurate method than PSA level, DRE or trans-rectal ultrasound-guided biopsy in the detection and localization of prostate carcinoma [6].

Prostate MRI is indicated if cancer is suspected despite negative biopsy findings and for local staging in patients confirmed with prostate cancer. Multiparametric scan technique is used in current routine prostate MRI scan, including T2W, DWI/ADC, dynamic contrast enhanced and spectroscopy. DWI MRI technique has increasingly been used in the detection and evaluation of prostate in these patients. Our study showed DWI/ADC MRI has a low sensitivity (52.14%) but has a very high specificity (93.48%) in detecting prostate cancer. DWI technique can be concluded as a very specific test and can be used as a diagnostic test for the detection of prostate cancer.

Table 1
Sensitivity and specificity

Performance measures	DWI $+$ ADC	DWI alone
Sensitivity	52.14%	25.00%
Specificity	93.48%	97.80%
Positive predictive value	77.66%	83.33%
Negative predictiv value	82.69%	75.00%
Accuracy	80.95%	75.76%

Comparison between sensitivity and specificity of DWI $+$ ADC and DWI alone in detection of prostate cancer ( $n=$ 462 sections).

Table 2

Test statistics

		Biopsy
		Cancer ( $+$ ve)	Cancer ( $-$ ve)
MRI	Cancer ( $+$ ve)	73 (TP)	21 (FP)
	Cancer ( $-$ ve)	67 (FN)	301 (TN)
	Total	140	322

Number of patients with positive and negative MRI and biopsy results based on prostate sections. TP: true positive, FP: false positive, FN: false negative, TN: true negative.

Table 3

Spearman’s rho correlation test

			ADC in malignant section	Gleason
Spearman’s rho	ADC in malignant section	Correlation coefficient	1.000	$-$ 0.244 ${}^{**}$
		Sig. (2-tailed)		0.004
		$N$	139	139
	Gleason score	Correlation coefficient	$-$ 0.244 ${}^{**}$	1.000
		Sig. (2-tailed)	0.004
		$N$	139	139

**: Correlation is significant at 0.01 level (2-tailed).

Table 4

Group statistics

	Biopsy	N	Mean	Std. deviation	Std. error mean
ADC	Benign	322	1554.0497	336.90983	18.77525
	Malignant	140	1092.3357	442.39566	37.38926

		Levene’s test
for equality
of variances	$t$ -test for equality of means
		$F$	Sig.	$t$	df	Sig. (2-tailed)	Mean difference	Std. error difference
95% confidence interval
of the difference
									Lower	Upper
ADC	Equal variances assumed	30.148	0.000	12.262	460	0.000	461.71398	37.65447	387.71789	535.71006
	Equal variances not assumed			11.036	212.099	0.000	461.71398	41.83858	379.24128	544.18667

Student $t$ -test showing mean ADC value, standard deviation and significant correlation between benign and malignant sections.

4.1 Diffusion weighted imaging (DWI)

DWI assesses the movement (Brownian motion) of free water in tissue. In tumors, the motion of water is restricted, probably due to their higher cellular density and increased nucleocytoplasmic ratio, and it can be depicted as bright signal on DWI image [7]. DWI alone is specific for detection of prostate cancer, but less sensitive than DWI $+$ ADC combined. By DWI alone, there is higher rate of false negative result compared to DWI combined with ADC. This might be due to T2 shine through artefacts, or signal from benign prostatic hyperplasia that can obscure the bright tumour focus on DWI. Because of these reasons, radiologist tends to miss a lesion by DWI interpretation alone. The sensitivity is increased when DWI is read in conjunction with ADC (52.14%) as compared when interpretation is based on DWI alone (25%) (Table 1). Other study has shown that the sensitivity and specificity of detection and localization of tumor is increased when DWI is combined with T2W [8].

4.2 Apparent diffusion coefficient (ADC)

From our observation, DWI in conjunction with visual analysis from ADC sequence shows a better sensitivity compared to DWI alone in detection of prostate tumour focus. The sensitivity is increased to 52.14% compared to DWI alone (25.00%). This is likely due to better contrast of low signal cancer focus on ADC against high signal of normal peripheral gland tissue. Quantitative ADC analysis allows the radiologist to choose an adequate cut off value to achieve a higher specificity and sensitivity [9]. From Student t-test statistical analysis, there is a significant difference between ADC values of benign and malignant tissue ( $p=$ 0.000) (Table 4). From ROC curve plotted for ADC values of benign and malignant sections (Fig. 5), we found that the cut off value of ADC between benign and malignant is at 1174.5 mm ${}^{2}$ /s at 93.48% specificity. The ROC curve shows a significant fair-to-good test with areas under the curve value of 0.79. It can be concluded that ADC value of lower than 1174.5 mm ${}^{2}$ /s can be taken as indicator of malignant tissue and values higher as benign. However, other resent study by Pietro et al. [10] has showed a cut off ADC value of $0.747\times 10^{-3}$ mm ${}^{2}$ /s was significantly correlated with the presence of aggressive tumour (Gleson score $\geqslant$ 7), which is lower than our finding [10].

Figure 1.

Example of high-grade tumour. MRI of a 77-year-old patient with prostate carcinoma (Gleason 8). The tumor focus in the peripheral zone of left apex shows bright signal on DWI (a) and low signal on ADC (b).

Figure 2.

Example of low-grade tumour. MRI of a 76-year-old patient (Gleason 6). The tumor in peripheral zone of left apex show slight increase in signal intensity compared to the rest of peripheral gland.

Figure 3.

False positive result. MRI of a 72-year-old man with BPH and raising PSA. MRI images in T2W (a), DWI (b), ADC (c) and contrast enhanced (d), show well defined hypointense lesion (arrow) at the junction of peripheral and central gland in left midzone, which shows restricted diffusion and contrast enhancement. However, biopsy came back as benign hyperplasia.

Figure 4.

False negative result. MRI of 74-year-old man with prostate carcinoma (Gleason 4) for local staging. a. ADC, b. DWI, c. T2W, d. post contrast. MRI shows hypertrophy of central gland with nodularity within suggestive of BPH changes. No definite T2W hypointense foci showing restricted diffusion or enhancement in the peripheral gland.

Figure 5.

Receiver operating characteristic (ROC) curve for ADC values.

Our study showed that ADC value of benign lesions has a consistent pattern distributing between 1000 mm ${}^{2}$ /s to 2500 mm ${}^{2}$ /s, with mean $\pm$ standard deviation of 1554.05 $\pm$ 336.91 ( $p=$ 0.000). This findings is expected as benign tissue should not show fluid restriction, thus has high ADC value. A few of the benign sections show low ADC value, which was also observed by a recent study by Rajan et al. [11]. Their study concluded that benign central zone tumour shows lower ADC value than other zones of prostate and overlap with the ADC value of tumor tissue [11]. However, ADC value of malignant lesions is more widely distributed between 500 mm ${}^{2}$ /s to 2000 mm ${}^{2}$ /s, with mean $\pm$ standard deviation of 1093.33 $\pm$ 442.4 ( $p=$ 0.000). This wide range of distribution results in overlapping between benign and malignant ADC values. For this reason, it is difficult for radiologists to determine the exact standardized range of ADC value for malignant lesion. Therefore, visual analysis from ADC sequence might be more helpful than cut off value in detecting prostate cancer. To use MRI as a modality to localize tumor focus as well as a screening and guiding tool for biopsy, it should have a high sensitivity and negative predictive value [12]. Although our study shows DWI/ADC has low sensitivity in detecting prostate cancer, it has a high negative predictive value (82.69%). This means that DWI/ADC technique used in our center, although less sensitive, is very specific and is useful as a diagnostic tool in screening and detecting tumor in patients with negative biopsy result but raising PSA values, local staging or post treatment follow up. It also has a potential role to be used as a guiding tool for repeat biopsy in patients with raising PSA values but has previous negative biopsy result.

4.3 ADC map for targeted biopsy tool

Decision on the management of prostate cancer has been difficult in high-risk patient who has negative TRUS biopsy result but have elevated PSA value [13]. Our study showed that DWI alone has a low sensitivity in detecting tumour focus. However, DWI with corresponding ADC has higher sensitivity to detect tumour focus. ADC map would be helpful together with DWI in localizing prostate carcinoma and guide for localization for subsequent biopsy. Our study found that ADC sequence increases the sensitivity in detection of prostate cancer when read in conjunction with DWI. This finding is supported by a recent study by Andrew et al. showing similar findings [14]. The visual analysis and quantitative analysis from ADC sequence helps in detection of tumour focus in prostate carcinoma as ADC shows low signal intensity on the background of high signal intensity of normal tissue, thus giving good contrast for tumour detection. A careful attention on ADC map would be helpful for better detection of prostate cancer when reporting prostate MRI. ADC value of 1174.5 mm ${}^{2}$ /s may be taken as cuff off value to distinguish between malignant and benign lesion. As ADC map improves detection of tumour focus on DWI MRI sequence, these technique used in conjunction may help in localizing high grade tumor focus for biopsy in patients with suspected prostate cancer. More targeted biopsy and sample cores can be obtained at areas showing restricted diffusion with high malignant potential, possibly reducing biopsy cores from other sections deemed having less malignant potential. A more structured study is recommended in the future using histopathological examination from prostate surgery for better accuracy of DWI and ADC sequence evaluation on detecting prostate carcinoma. Labeled section by section analysis of the prostate gland on histopathological specimens should corresponds to the section as seen on MRI for accurate results. Study that look into the acquisition, processing and interpretation of prostate MRI sequences, particularly DWI and ADC sequences, is also helpful so that a more standardized imaging protocol could be implemented for better detection of tumour focus in prostate MRI.

5. Conclusion

DWI technique has a low sensitivity in detecting tumour in prostate cancer, likely because it is a technique sensitive/prone to susceptibilty artefacts and signal distortion by haemorrhage, inflammation/prostatitis or movement artefacts. ADC showed better detection of tumor as it provides more contrast between tumor focus and surrounding tissue. The visual and quantitative analysis from ADC sequence helps in detection of tumour focus in prostate carcinoma as ADC shows low signal intensity on the background of high signal intensity of normal tissue, thus giving good contrast for tumour detection. A careful attention on ADC map would be helpful for better detection of prostate cancer when reporting prostate MRI. Based on the finding from this study, ADC value of 1174.5 mm ${}^{2}$ /s may be taken as cuff off value to distinguish between malignant and benign lesion, though there is significant overlap between benign and malignant tissue ADC values. There is a negative correlation between ADC values and Gleason score. This finding indicate that DWI and ADC has a potential role to determine aggressiveness of tumour. With high specificity (95.34%) and negative predictive value (82.69%) from this study, DWI/ADC technique used in our center is useful as a diagnostic tool in screening and detecting prostate cancer and could be used as a guiding tool for prostate biopsy in the future.

Footnotes

Acknowledgments

I wish to acknowledge Prof Dr. Syed Zulkifli Syed Zakaria from Department of Paediatrics, Universiti Kebangsaan Malaysia Medical Center for his guidance on statistical analysis of this study.

References

Hoeks

C.M.

Barentsz

J.O.

Hambrock

, et al. Prostate cancer: Multiparametric Mr imaging for detection, localization, and staging, Radiology 261 (2011), 46–66.

Koh

D.M.

et al., Diffusion weighted MRI in the body: Applications and challenges in oncology, AJR 188 (2007), 1622–1635.

Sally

E.E.

Manal

Lamiaa

I.A.M.

and Rania

M.A.A.

, Diffusion-weighted MR imaging and ADC measurement in normal prostate, benign prostatic hyperplasia and prostate carcinoma, The Egyptian Journal of Radiology and Nuclear Medicine 45(2) (2014), 535–542.

Catalona

W.J.

Partin

A.W.

Slawin

K.M.

Brawer

M.K.

Flanigan

R.C.

Patel

Richie

J.P.

Dekernion

J.B.

Walsh

P.C.

Scardino

P.T.

Lange

P.H.

Subong

E.N.

Parson

R.E.

Gasior

G.H.

Loveland

K.G.

and Southwick

P.C.

, Use of the percentage of free prostatespecific antigen to enhance differentiation of prostate cancer from benign prostatic disease: A prospective multicenter clinical trial, J Am Med Assoc 279 (1998), 1542–1547.

Tchetgen

M.B.

and Oesterling

J.E.

, The effect of prostatitis, urinary retention, ejaculation, and ambulation on the serum prostate-specific antigen concentration, Urol Clin North Am 24 (1997), 283–291.

Mullerad

Hricak

Kuroiwa

Pucar

Chen

H.N.

Kattan

M.W.

and Scardino

P.T.

, Comparison of endorectal magnetic resonance imaging, guided prostate biopsy and digital rectal examination in the preoperative anatomical localization of prostate cancer, J Urol 174 (2005), 2158–2163.

Tan

Wei

Johnson

and Kundra

, Diffusion-weighted MRI in the detection of prostate cancer: Meta analysisCher, AJR 199 (2012) 822–829.

Noha

M.A.

Hytham

H.E.

and Essam

A.A.

, The role of diffusion – weighted MRI in evaluation of prostate cancer, The Egyptian Journal of Radiology and Nuclear Medicine 45(1) (2014), 231–236.

Ozgür

Tan

, Ercannci Arda

Sibel

Neslihan

and Nevzat

, Diffusion-weighted MRI of urinary bladder and prostate cancers, Diagn Interv Radiol 15 (2009), 104–110.

10.

Pietro

Davide

Antonio

, et al., Multiparametric MRI apparent diffusion coefficient (ADC) accuracy in diagnosing clinically significant prostate cancer, In Vivo 31(3) (2017), 415–418.

11.

Rajan

T.G.

Christopher

R.K.

Kirema

G.R.

et al., Apparent diffusion coefficient values of the benign central zone of the prostate: Comparison with low- and high grade prostate cancer, AJR 205(2) (2015), 331–336.

12.

Myung

S.C.

Yong

S.C.

Byung

I.Y.

J.K.

Hyuk

J.C.

Sung

H.H.

Y.L.

Tae-Kon

and Sae

W.K.

, The clinical value of performing an MRI before prostate biopsy, Korean J Urol 52(8) (2011), 572–577.

13.

KiM

C.K.

Park

B.K.

and Kim

, Diffusion-weighted MRI at 3T for the evaluation of prostate cancer, AJR 194 (2010), 1461–1469.

14.

Andrew

B.R.

, et al., Prostate cancer: Comparison of tumor visibility on trace diffusion-weighted images and the apparent diffusion coefficient map, AJR 196 (2011), 123–129.