Background

A common malignancy among women is breast cancer (BC). Using DNA microarray techniques, the gene pattern of expression in breast cancer has been classified into different molecular types with clinical, biologic, and treatment effects based on estrogen (ER), progesterone (PR) and HER-2 receptor analysis. Breast cancer has four primary molecular subtypes, defined mainly by hormone receptors (HR) and other types of proteins involved (or not) in each cancer: Luminal A or HR + /HER2- (HR-positive/HER2-negative); Luminal B or HR + /HER2 + (HR-positive/HER2-positive); triple negative or HR-/HER2- (HR/HER2-negative); and HER2-positive [1].

Ki-67 is a nuclear protein related to cellular proliferation and was first identified by Gerdes et al. in the early 1980s [2]. Ki-67 is found in all proliferating cells and is used as a proliferation marker; it is also considered a prognostic factor for breast cancer [3, 4]. Compared with other markers, Ki- 67 immuno-staining is a suitable method for testing proliferating index [5, 6].

Various imaging modalities including mammography, ultrasound, and magnetic resonance (MRI) play a vital role in diagnosing and staging breast cancer. Imaging cannot only detect breast tumors but can also predict their histopathological characteristics, yet none of these imaging modalities can provide accurate information about tissue cellularity, which is considered the main indicator of tumor grade. Therefore, many investigators try to assess by using diffusion-weighted breast imaging and without intravenous contrast material injection, diffusion-weighted breast imaging can facilitate the diagnosis of breast lesions, yet may require some increase in examination time [7, 8, and 9].

Diffusion-weighted imaging of the breast is an MRI sequence that gives several advantages compared to dynamic contrast-enhanced sequence as it does not require intravenous contrast, short duration, and easy to perform. Moreover, integrating quantitative imaging biomarkers gives us the ability to measure the underlying pathological mechanisms in vivo noninvasively, identifying properties that are important for detection, diagnosis, prognosis, or response to therapy [10].

In this study, we aimed to assess the status of apparent diffusion coefficient (ADC) in breast cancer patients and to evaluate its association with other factors including Ki- 67, molecular subtype, and grade of breast tumor to validate its significance as a prognostic factor of breast cancer.

Methods

This was a retrospective study. It was approved by the institutional review board. Between April 2016 and February 2020, the total number of included patients was 55.

Inclusion criteria: (1) patients with breast malignancy proved by pathology either by surgery or by biopsy; (2) patients who had a standard MRI breast study, including axial T1WI, fat-suppressed T2 WI, axial fat-saturated T1WI pre- and post-enhancement, and DWI sequences; and (3) patients with complete clinical data; immunohistochemistry and Ki-67 values in their histopathology reports. Exclusion criteria: (1) operated/ treated breast before MRI; (2) poor image quality; and (3) incomplete immunohistochemistry or Ki-67 values in their reports.

MRI technique

DCE-MRI breasts were performed on a 1.5-T system (MAGNETOM Aera; Siemens, Erlangen, Germany). Women lying prone with their breasts fixed in a dedicated four-channel phased-array breast coil. All written consents were obtained before the study was performed. Patients were advised not to move to avoid motion artifacts.

The following MRI sequences were taken: slice thickness 4 mm, FOV 350 mm, and matrix 512 × 512: Axial Turbo Inversion Recovery Magnitude (TIRM) with TR 7700 ms and TE 74 ms; Axial T2 fast spin-echo (FSE) with TR 6160 ms and TE 76 ms; and Axial T1 fast spin-echo (FSE) with TR 415 ms and TE 4.6 ms. Diffusion-weighted imaging (DWI) was done at TR 8200 ms and TR 85 ms in the axial plane bilaterally with b values of 0, 500, and 1000 s/mm2. ADC map was systematically performed. ADC value was measured in mm2/s using a rounded ROI with a diameter of 5–10 mm2 placed on the darkest point of the mass which shows diffusion restriction on the DW sequences. Areas of T2 shine-through, such as cystic or necrotic portions of the tumor shown as high signal intensity on T2-weighted images and ADC maps, were avoided. When comparing with DCE MR images, the enhancing solid portion of the mass was used to locate ADC measurements. The ADC value was automatically calculated when the ROI was drawn. Two radiologists with 20 and 25 years of experience in MRI breast analyzed the images. The diagnosis was reached by consensus.

Histological analysis

According to the numerical scoring pattern to define tubule formation, pleomorphism, and mitotic count, Nottingham combined grading was used to assess histological grades of invasive ductal carcinoma, NOS (non-otherwise specified). The score ranges from 3 to 9, with a total score of 3–5 representing low grade (grade 1), a score of 6 or 7 representing grade 2, and a score of 8 or 9 representing grade 3 (Table 1): Nottingham combined histological grade [11] (Figs. 1, 2, 3, 4).

Table 1 Nottingham combined histological grade, Ref [11]
Full size table

Statistical analysis

Data were coded and entered data using SPSS (Statistical Package for the Social Sciences) version 23. Data were summarized using the mean and standard deviation. For quantitative data, we used median, minimum, and maximum; for qualitative data, frequency (number) and relative frequency (%) were used for categories. A comparison of quantitative variables was made using nonparametric Kruskal–Wallis and Mann–Whitney tests [12]. ROC curve was plotted with an area under the curve analyzed to detect the specific cutoff value of ADC for the detection of high-grade tumors. P values below 0.05 (< 5%) were considered statistically significant.

Results

Fifty-five female patients were included in this study, with an age range from 21 to 83 years (mean age 50.8 ± 14). The included carcinomas included 48 (87.3%) invasive ductal carcinomas; 26 (54%) of them were NST (no specific type), 3 (5.5%) mucinous carcinoma, 2 (3.6%) invasive lobular carcinoma, 1 (1.8%) NEC (neuroendocrine carcinoid), and 1 (1.8%) carcinosarcoma/metaplastic carcinoma.

Their grading was as follows: 11 lesions (20%) grade I, 38 lesions (69.1%) grade II, and 6 (10.9%) lesions grade III. Ten (18.2) lesions were molecular type A, 35 (63.6) lesions were molecular type B, 5 (9.1) were HER2 overexpression, and 5 (9.1) were triple negative (Table 2).

Table 2 Description of study variables
Full size table

Analysis of diffusion-weighted images and comparison of the mean ADC value as well as Ki-67 with the pathological grade and molecular subtype of the tumor were performed. In all the studied lesions, 55 (100%) showed a bright signal in DWI and low/intermediate signal in ADC map denoting diffusion restriction. ADC values ranged from 0.5 to 1.6 × 10–3 mm2/s (mean 1 ± 0.2 × 10-3mm2/s). The mean ADC value of grade I was 1.0 ± 0.2 × 10–3 mm2/s, grade II was 0.9 ± 0.1 × 10–3 mm2/s, and grade III was 0.9 ± 0.2 × 10–3 mm2/s (Table 3). There was no significant difference regarding the mean ADC values of tumors grade I, II, and III. The mean ADC value of molecular type A, molecular type B, HER2 overexpression, and triple-negative subtypes was 1.2 ± 0.2 × 10–3 mm2/s, 0.9 ± 0.1 × 10–3 mm2/s, 0.9 ± 0.1 × 10–3 mm2/s, and 0.8 ± 0.1 × 10–3 mm2/s, respectively. Yet a significant difference between the mean ADC value of tumors of different molecular subtypes (P = 0.000) was found. A significant difference between the mean ADC value of tumors of molecular type A as opposed to molecular type B (P = 0.000), HER2 overexpression (P = 0.018), and TN (P = 0.000), respectively, was found. However, there was no significant difference between molecular type B, HER2 overexpression, and TN (Tables 4, 5, 6).

Table 3 Comparison of ADC regarding grade
Full size table
Table 4 Comparison of ADC regarding molecular type
Full size table
Table 5 Pairwise comparison of ADC regarding molecular type
Full size table
Table 6 ROC curve analysis in differentiating molecular type A through ADC value
Full size table

Ki-67 values ranged from 6 to 92.5 with a mean of 42.9 ± 17.3; the mean Ki-67 of grade I, grade II, and grade III was 41.3 ± 18.2, 40.9 ± 16.2, and 58.8 ± 17.2, respectively. There was no significant difference between the Ki-67 value of tumors of grade I, II, and III (Table7).

Table 7 Comparison of Ki67 regarding grade
Full size table

The Ki-67 value of tumors of different molecular subtypes showed a significant difference (P = 0.000) (Table 8). A significant difference was found between the mean ADC value of tumors of molecular type A as opposed to molecular type B (P = 0.000), HER2 overexpression (P = 0.014), and TN (P = 0.000), respectively. However, there was no significant difference between the last three types together: molecular type B, HER2 overexpression, and TN (Tables 5 and 6).

Table 8 Comparison of Ki67 regarding molecular type
Full size table

Correlation between ADC and Ki-67 value showed an R value of -0.779 and R2 value of 0.599 (P = 0.000) (significant/good negative correlation) (Tables 9 and 10).

Table 9 Correlation between ADC and Ki67 value
Full size table
Table 10 ROC curve analysis in differentiating molecular type A from other molecular types through Ki67 value
Full size table

Discussion

Different molecular types of breast carcinoma show different biological characteristics, clinical outcomes, and prognoses [13]. An accurate assessment of the disease course and prognosis preoperatively is essential for accurate diagnosis planning for treating breast cancer [14].

The value of breast MRI in the preoperative assessment of patients is important owing to its value as a means of describing pattern, size, and number of breast lesions. The early proliferation of tumor accompanies changes in both anabolism and catabolism affecting its size and altering the intra- and extracellular environment. Functional MR as diffusion-weighted imaging (DWI) and apparent diffusion coefficient (ADC) assist the radiologist in finding certain changes related to tumor proliferation [15].

In the current study, the calculated mean ADC value of breast malignant masses measured 1 ± 0.2 × 10–3 mm2/s. This was similar to mean ADC values reported by previous studies, such as Belli et al. [13] and Costantini et al. [16], where mean ADC values measured 1.02 × 10–3 mm2/s and 1.03 × 10–3 mm2/s, respectively. However; other studies reported lower ADC including Kato et al. [17] who reported the mean ADC value of 0.894 ± 0.20 × 10–3 mm2/s; it was 0.85 ± 0.12 × 10–3 mm2/s in Gouhar et al. [18] study, 0.91 ± 0.20 × 10–3 mm2/s in Park et al. [19] study, 0.93 ± 0.27 × 10–3 mm2/s in Ulghaffara et al. [20] study, and 0.91 ± 0.151 × 10–3 mm2/s in Matsubayashi et al. [21] study.

ADC values ranged from 0.5 to 1.6 × 10–3 mm2/s in our study, thus lying above the cutoff value differentiating benign from malignant lesions suggested by previous studies, such as Sharma et al. [15] with a cutoff ADC value of 1.23 × 10−3 mm2/s, and Tan et al. [22] where the cutoff ADC value for benign lesions was 1.21 × 10−3 mm2/s for b = 500 s/mm2 and for malignant lesions was 1.22 × 10−3 mm2/s for b value:1000 s/mm2. Studying the correlation between the mean ADC value and histological grades of breast carcinoma, we found that there was no significant relation between them.

This coincided with the previous research done by Park et al. [19] and Tan et al. [22] who found no significant correlation between the measured ADC values and tumor grades.

On the contrary, Belli et al. [13], Abdel Razek et al. [23], Costantini et al. [16], Gouhar et al. [18], and Guo Yan et Al. [24] found that there was a significant yet inverse relation between them.

In the current study, the mean ADC value for grade I masses was 1.0 ± 0.2 × 10–3 mm2/s, that of grade II was 0.9 ± 0.1 × 10–3 mm2/s, and that of grade III was 0.9 ± 0.2 × 10–3 mm2/s. Kindly view Figs. 1, 2, 3 and 4 (case presentations). 

Fig. 1
figure 1

A 42-year-old female with multifocal left breast ca A MLO low-energy CESM (B), (C) recombined CESM MLO and CC showing multiple enhancing masses lower inner quadrant D DCE-MRI shows marginal enhancement with central necrosis E and F DWI b value = 800 and ADC, respectively, showing peripheral diffusion restriction. ADC value = 0.989 × 10−3mm2/s. HER-2 positive BC. IDC GRADE III

Full size image
Fig. 2
figure 2

A 52-year-old female with recently diagnosed left breast LOQ IDC Grade II (Luminal type A) A MRI Wash-in map B, C DWI b value = 800 and ADC, respectively, showing diffusion restriction ADC value = 0.123 × 10−3mm2/s D 3D Tomo slice CC view a single spiculate mass (arrow) E DCE-MRI sagittal shows intense marginal enhancement

Full size image
Fig. 3
figure 3

A 49-year-old female with bifocal right breast IDC Grade II (luminal subtype B) A DWI b value:800 B ADC showing diffusion restriction ADC value = 0.994 × 10−3mm2/s. C DCE-MRI showing intense enhancement. D 3D MIP E delayed sagittal MRI showing central wash-out with marginal enhancement. Ki-67 high 25%

Full size image
Fig. 4
figure 4

A 72-year-old female with an incidentally discovered single circumscribed left breast mass with homogenous enhancement as seen on mammogram CC view (A) and recombined CESM (B). C and D DWI b value = 800 and ADC, respectively, showing diffusion restriction. ADC value = 0.789 × 10−3mm2/s E Axial MRI Left breast shows hypointense mass with smooth margin F DCE-MRI 3D MIP shows the left breast mass with feeding vessels (arrow) G US shows the mass with circumscribed margin, posterior enhancement and advancing anterior edge. Note peripheral vascularity by color Doppler. Core biopsy revealed metastatic NEC, triple negative. H Axial CT scan shows the primary mesenteric carcinoid tumor with calcifications

Full size image

This is equal to the results of Costantini et al. [16] reporting mean ADC values for grade I, II, and III tumors to be: 1.25 × 10−3 mm2/s, 1.02 × 10−3 mm2/s, and 0.92 × 10−3 mm2/s, respectively. However, it differs from the results obtained by Gouhar et al. [18] who reported that the mean ADC values of grades I, II, and III were 0.96 ± 0.12 × 10–3 mm2/s, 0.87 ± 0.07 × 10–3 mm2/s, and 0.75 ± 0.12 × 10–3 mm2/s, respectively. This difference may be attributed to the discrepancy in patient number and diffusion technique using different b values.

High ADC values can be useful in diagnosing TNBC [25]. In our study, we had 5 cases of TNBC with a mean ADC value measuring 0.8 × 10−3mm2/s. It is postulated that TNBC shows higher ADC values due to necrotic tissue.

In previous studies, breast carcinomas showing high expression of Ki 67 had lower ADC values as opposed to tumors with low Ki 67 expression (26). In breast cancer, an insignificant correlation between both was found. Thus, ADC cannot be used as a proliferation marker in breast cancer [27]. This differs from our results which showed a good negative correlation between both ADC and Ki-67. Low ADC values were associated with tumors with high Ki-67 expression which is a predictor of proliferation and prognosis. This difference may be attributed to a limited number of patients, uneven distribution of the variable histological grades, and molecular subtypes.

Limitations of this study

The limited number of patients might have affected our results, especially that all our patients were pathologically proven malignant breast lesions. No cutoff value for luminal subtypes has been elicited during our study result analysis; this might be due to overlap between readings among molecular subtypes. Further studies on a larger scale of patients should be encouraged.

Conclusions

There was a significant inverse correlation between ADC values and Ki-67 expression. ADC values could be a good discriminator between tumors of molecular subtype A from other subtypes. Ki 67 could also be a good discriminator between tumors of molecular subtype A from other subtypes. Both ADC values and Ki- 67 did not show a correlation with the pathological grade of the tumor.