Apparent diffusion coefficient value as a biomarker reflecting morphological and biological features of prostate cancer
- First Online:
- Cite this article as:
- Bae, H., Yoshida, S., Matsuoka, Y. et al. Int Urol Nephrol (2014) 46: 555. doi:10.1007/s11255-013-0557-1
- 215 Views
To assess whether there is an association between the apparent diffusion coefficient (ADC) value and the pathological characteristics of prostate cancer.
The study cohort consisted of 29 consecutive patients with prostate cancer treated with radical prostatectomy. All patients underwent diffusion-weighted MRI before the prostate biopsy. In 42 tumor foci, the associations of the ADC values with the clinicopathological characteristics and Ki-67 labeling index (LI) were analyzed.
High-grade cancers (Gleason score [GS] ≥ 4 + 3), larger cancers (maximum diameter (MD) ≥ 16 mm), and highly proliferating cancers (Ki-67 LI ≥ 4.43 %) had significantly lower ADC values, respectively (P < .001, P = .008, and P = .044, respectively). There was no significant difference in ADC value according to age, prostate-specific antigen, presence of extra-prostatic extension, and intra-tumoral stroma proportion. Multivariate analysis showed that GS, Ki-67 LI, and MD had independent and significant correlations with ADC value (P < .001, P = .006, and P = .002, respectively). Low ADC tumors (<0.52 × 10−3 mm2/s) are likely to be high-grade cancer foci compared with high ADC tumors (relative risk: 65.2). The sensitivity and specificity of the ADC value to predict high-grade cancer foci are 81.8 and 93.5 %, respectively.
A low ADC value reflects the morphological and biological features of prostate cancer. Analyzing the ADC value may make it possible to more precisely predict the cancer aggressiveness of each focus before treatment.
KeywordsDiffusion magnetic resonance imagingProstatic neoplasmsBiological markersKi-67 Antigen
The treatment strategy of localized prostate cancer is diverse ranging from active surveillance to radical whole-gland therapy. Recently, as an individualized and optimized treatment option, focal treatment against prostate cancer has received a great deal of attention in regard to its ability to selectively eradicate significant cancer focus while preserving uninvolved parenchyma [1, 2]. Considering the biological heterogeneity and multiplicity of prostate cancer, evaluating the biological character of each prostate cancer focus precisely is essential in the selection of the tumor foci that need to be treated or untreated. Therefore, it is crucial to predict the Gleason score (GS) of each tumor foci precisely.
To date, T2-weighted MRI has been commonly used to detect prostate cancer. Recently, diffusion-weighted MRI (DW-MRI) has been widely introduced in the clinical setting. It is advantageous as it offers increased diagnostic accuracy due to the clear delineation between normal and prostate cancer, namely the high signal of cancerous lesions and the restricted signal of normal tissue [3–5]. DW-MRI is a non-invasive imaging technique that quantifies the diffusion of water molecules in tissues without any contrast agents, tracers, or exposure to radiation . DW-MRI may also provide qualitative information regarding the pathophysiological character of prostate cancer . We have reported the utility of multi-parametric MRI including DW-MRI to exclude the significant prostate cancer in low-risk patients . The degree of diffusion is quantitatively measured as the apparent diffusion coefficient (ADC) value. Some previous papers showed the correlation between ADC value and the feasibility of ADC value to predict cancer aggressiveness [8, 9].
To support the use of DW-MRI in the qualitative assessment of prostate cancer, it is important to link the ADC value to underlying pathophysiological characteristics. In the current study, the relationship between ADC value and multiple biomarkers, including the Ki67-labeling index (LI) (a marker of cell proliferation), and stroma to cancer cell ratio as well as GS on prostatectomy were assessed. Furthermore, to evaluate the clinical benefit of assessing ADC values, we analyzed the role of ADC value for predicting GS on prostatectomy.
Materials and methods
An institutional review board approved this study, and informed consent was obtained from all participants. In our institute, as a prospective study, pre-biopsy multi-sequence MRI including DW-MRI has been performed in all patients with a suspicious of prostate cancer. Of these, 29 consecutive patients were treated with radical prostatectomy under the diagnosis of prostate cancer by biopsy in our institute between November 2009 and January 2011. No other medical treatment for prostate cancer was performed prior to the radical prostatectomy. We retrospectively evaluated these 29 patients in this study.
Multi-sequence MRI, including T1- and T2-weighted imaging and DW-MRI, was performed using a 1.5-Tesla imager (Intera Achieva; Philips, Best, Netherlands) with a 32-channel sensitivity encoding body coil under free breathing. The imaging parameters of DW-MRI with a single-shot spin-echo planar imaging sequence were set as follows: repetition time, 5,000 ms; echo time, 80 ms; matrix, 128 × 99; field of view, 300 × 255 mm; slice thickness, 4 mm; interslice gap, 0.4 mm; three different diffusion gradient b values (b = 0, 1,000, 2,000 s/mm2); fat suppression, spectral pre-saturation inversion recovery.
Pathologists evaluated prostatectomy specimens and determined GS and pathological T stage in accordance with the 2005 International Society of Urologic Pathology modified Gleason system and the 2009 TNM system, respectively. The maximum diameter (MD) of tumor foci was also determined from the prostatectomy specimen. In patients with multiple tumor foci, only the largest two tumors, the MD of which was larger than 5 mm, were evaluated in this study. The intra-tumoral stroma proportion was evaluated by visually estimating the percentage of stroma area in the tumor at 40× magnification of hematoxylin and eosin-stained sections.
ADC value measurement
Transverse ADC maps were constructed using DW-MRI data at a workstation (View Forum R4.1, Philips Healthcare). After reviewing the location of prostate cancer foci on prostatectomy specimens, one urologist with 15 years of experience in treating prostate cancer and 5 years of experience in reading DW-MRI data of prostate cancer analyzed the ADC value of the prostate cancer foci of the corresponding MRI cuts. The regions of interest (ROIs) were retrospectively located to cover the lesions corresponding with the tumor foci with reference to histological findings (Fig. 1), and the ADC values of each ROI were calculated by applying the following formula: ADC = −ln(S/S0)/(b − b0), where S0 and S are the signal intensities obtained with two different diffusion gradient values (b0 and b; 1,000 and 2,000 s/mm2, respectively). The minimum ADC value within the ROI was applied as an index parameter for the diffusion of the prostate cancer focus.
Spearman’s rank correlation test was used to assess the correlation among ADC value, Ki-67 LI, and MD. GS was divided into two groups (GS ≥ 4 + 3: high-grade cancer vs. GS ≤ 3 + 4: low-grade cancer). Age, prostate-specific antigen (PSA), MD, Ki-67 LI, and intra-tumoral stroma proportion were divided at the median value of each parameter. The difference in continuous variables between the two groups was evaluated by the Wilcoxon test. Partition analysis was used to set the best cutoff for ADC value to predict the presence of high-grade cancer. The statistical analyses were performed using JMP ver. 7.0 (SAS Institute) and R ver. 2.15.2 (R Core Team, 2012, R Foundation for Statistical Computing, Vienna, Austria). A P value of <.05 was considered statistically significant.
Characteristics of tumors
3 + 3/3 + 4
4 + 3/4 + 4/4 + 5
Maximum diameter (mm)
ADC value (×10−3 mm2/s)
Ki-67 LI (%)
Intra-tumoral stroma proportion (%)
ADC value according to clinical and pathological parameters
ADC value (×10−3 mm2/s)
≤3 + 4
≥4 + 3
Ki-67 LI (%)
Intra-tumoral stroma proportion (%)
Maximum diameter (mm)
Furthermore, we evaluated the correlation among ADC value, Ki-67 LI, and MD. Ki67-LI and MD were significantly and inversely correlated with ADC value (ρ = −0.40, P = .009 and ρ = −0.42, P = .006, respectively), whereas Ki-67 LI and MD did not appear correlated.
Regression analysis for predicting ADC value
Full model P value
Reduced model P value
≥67 versus <67
≥6.5 versus <6.5
≥4 + 3 versus ≤3 + 4
Ki-67 LI (%)
≥4.43 versus <4.43
Intra-tumoral stroma proportion (%)
≥40 versus <40
Positive versus negative
Maximum diameter (mm)
≥16 versus <16
Lastly, we analyzed whether ADC value of the tumor focus can be a predictor of high-grade cancer. In the current tumor-based analysis, logistic regression analysis showed that the ADC value was a significant predictor of high-grade cancer focus. At the cutoff ADC value of 0.52 × 10−3 mm2/s, 11 (26.2 %) and 31 patients (73.8 %) were classified as having low and high ADC tumors, respectively. Low ADC tumors are likely to be high-grade cancer foci compared with high ADC tumors (relative risk: 65.2). The sensitivity and specificity of the ADC value to predict high-grade cancer foci were 81.8 and 93.5 %, respectively.
The current study demonstrated the significant associations of the ADC value with GS as well as Ki-67 LI and MD in prostate cancer. These results suggest that, given the reported prognostic value of all three parameters, the ADC value has potential to provide clinically important qualitative information on prostate cancer. This study also presented preliminary data that suggested the feasibility of detecting the high-grade cancer foci. This MRI finding-based qualitative analysis of tumor focus might make it possible to characterize the aggressiveness of each tumor in prostate. When we consider the focal treatment to only significant cancer focus as an individualized treatment, ADC value could be a useful biomarker, although confirmatory studies are needed.
DW-MRI is a functional imaging technique that is constructed by quantifying the diffusion of water molecules in tissues. The benefit of applying DW-MRI for detecting various types of cancer has been reported . The ADC value is a quantitative parameter of the extent of water molecule diffusion. In this unique measure, the ADC value of cancerous tissue is lower than the surrounding tissue, due to its higher cellularity, tissue disorganization, and decreased extracellular space, all of which restrict water diffusion. This is because the cellular and structural changes of cancer cells impede water diffusion within the tissues, which affects the ADC value. Because DW-MRI provides pathophysiologic information on tissues, the ADC value could potentially be employed to predict the biological aggressiveness. However, the mechanism of this characteristic of the ADC value is not yet fully understood.
The Gleason system is the most widely accepted histological grading system based upon the glandular pattern of the tumor. Cytologic features are not considered in this grading system. Consistent with the reports by Woodfield et al. and Oto et al., our results showed that the ADC values of high-grade cancer tend to be lower than those of low-grade cancer [11, 12]. Considering that the GS is classified depending on the degree of loss of the normal glandular structure, including shape, size, and differentiation of the glands, low-grade cancer has low tumor cellularity and larger extracellular spaces compared with high-grade cancer. The morphological change of the tumor glands may be the underlying mechanism for the association between the ADC value and GS in prostate cancer. We also hypothesized that the intra-tumoral stroma proportion has some impact on ADC value, according to the report from Aoyagi et al. . However, in the current study, we could not found any relationship between the ADC value and the intra-tumoral stroma proportion.
A significant association of the ADC value with cell proliferation can be observed in various malignancies. In urothelial cancer cells of the bladder and upper urinary tract, we reported the significant correlation between the ADC value and Ki-67 LI [13, 14]. In the case of prostate cancer, Zelhof et al.  reported the correlation of cell density with the ADC value as a cancer characteristic that reflects cell proliferation. The ADC value obtained in the current study was independently and significantly correlated with the proliferative activity assessed by Ki-67 LI. To our knowledge, this is the first study to show the relationship between the ADC value and Ki-67 expression in prostate cancer. The morphological changes secondary to the higher proliferative activity of tumor cells may explain this correlation. Our result also showed that MD was an independent factor of the ADC value. Considering our result showing high-grade cancers were tend to be significantly larger than low-grade cancers (P = .034, data not shown), the biological character of aggressive tumors expressed as MD and Ki-67 LI influences the ADC value of prostate cancers.
The current study proposes a cutoff of the ADC value to predict high-grade cancer foci. Despite the considerable overlap between ADC values of low-grade and high-grade cancer, the sensitivity and specificity for detecting high-grade cancer were 81.8 and 93.5 %, respectively. This diagnostic procedure may make it possible to perform tumor foci-based grading risk stratification when selective focal treatment against significant cancer lesions is being considered. Recently, excellent detectability of prostate cancer in T2-weighted images has been introduced as an individualized prostate biopsy technique, which is part of an MRI-ultrasound fusion prostate biopsy. Some papers have reported improved prostate cancer detection rates by targeting biopsy to MRI-positive foci using a fusion of MRI and transrectal ultrasound [16, 17]. Incorporating the GS on targeted biopsy with the ADC value will further improve the detection of high-grade cancer foci.
ADC value measurement has not yet been standardized. In some papers, the minimum ADC value within the ROI was used to evaluate the correlation with the biological aggressiveness of the tumor [9, 18–20], whereas the mean ADC value was used in other papers . We used the mean ADC value in evaluating urothelial cancer in the bladder and upper urinary tract because of the intratumoral homogeneity of the ADC value within the ROI [13, 21–24]. However, because of the heterogeneity of the intratumoral ADC value in prostate cancer, we opted to use the minimum ADC value within the ROI as an index of the diffusion environment rather than the mean ADC value, although GS, Ki-67 LI, and MD also significantly affected the mean ADC value of prostate cancer (P < .001, P = .034, and P < .001, respectively, data not shown). Moreover, the mean ADC value was also a significant predictor of high-grade cancer foci (P = .0092, data not shown).
The present study has some limitations. First, the study had a small sample size and was retrospective in nature. A study with a prospective larger cohort is needed to confirm the current findings. Furthermore, we could not validate the suggested cutoff ADC value for predicting high-grade cancer foci on prostatectomy; validation of the results is needed with another data set. Second, we have to consider the partial volume effect in evaluating the ADC value of small tumors, especially in the 5 tumors smaller than 10 mm in MD, because the slice thickness of the DW-MRI image was 4 mm. However, after excluding 5 tumors smaller than 10 mm in MD, ADC value of the remaining 37 rumors still had a significant association with Ki-67 LI and MD (ρ = −0.42, P = .011 and ρ = −0.36, P = .027, respectively). The result of this sub-analysis indicates that the partial volume effects in analyzing the ADC value of small tumors are unlikely to change the conclusion of this study. Third, the absolute cutoff of the ADC value applied in any situation cannot be intrinsically defined because various imaging protocols and imaging instruments are broadly used in clinical practice. The standardization of the ADC value measurement would be desirable. Finally, as an intrinsical limitation of this study in comparing imaging with histological findings, slice interval of histologic sections and MR images may lead to mismatching the slices. However, applying the minimum ADC value minimizes the effect of this limitation because the ADC values of the tumor are expected to be lower than those of the surrounding tissue.
A low ADC value reflects the morphological and biological features of prostate cancer. Analyzing the ADC value may enable the more precise prediction of the cancer aggressiveness of each focus before treatment.
Conflict of interest
The authors declare that they have no conflict of interest.