Soft tissue sarcoma: intravoxel incoherent motion and diffusion kurtosis imaging parameters correlate with the histological grade and Ki-67 expression

Abstract

Background

Accurate prediction of the histological grade and Ki-67 expression of soft tissue sarcoma (STS) before surgery is essential for the subsequent diagnosis, treatment, and prognostic evaluation of patients.

Purpose

To evaluate intravoxel incoherent motion (IVIM) and diffusion kurtosis imaging (DKI) in predicting the histological grade and Ki-67 expression of STS.

Material and Methods

A total of 40 patients underwent 3-T MRI, including conventional sequences; IVIM and DKI parameters were obtained. All patients were divided into a low-grade (grade 1 and grade 2) group and a high-grade (grade 3) group through pathological analysis. Ki-67 expression of each lesion was calculated. Chi-square test, independent sample t-test, Mann–Whitney U test, Pearson, Spearman, and receiver operating characteristic curve analysis were performed.

Results

There were 17 patients in the low-grade group and 23 in the high-grade group. Ki-67 expression was in the range of 10%–80%. D value was inversely correlated with Ki-67 expression. MK value showed a moderate positive correlation with Ki-67 expression. Regarding histological grading, only the peritumoral enhancement was statistically different between low- and high-grade STS on conventional MRI (P=0.024). The high-grade group had significantly higher MK value and lower D and MD value than the low-grade group. MK value showed the best diagnostic performance. The combination of MK and MD yielded the highest specificity (88.24%), and the combination of D, MK, and MD yielded the best area under the curve value (0.841) and sensitivity (95.65%).

Conclusion

IVIM and DKI parameters were correlated with Ki-67 expression and could help differentiate between low- and high-grade STS.

Keywords

Soft tissue sarcoma diffusion-weighted imaging histological grade

Introduction

Soft tissue sarcoma (STS) is a rare type of heterogeneous tumor. Studies had pointed out that histological grade was an independent and most important prognostic factor for STS (1 –3). At present, the most widely used was the scoring and grading system proposed by the FNCLCC system (4). In previous studies, scholars tended to classify grade 2 and grade 3 STS into the high-grade group (5,6). However, we noted that other studies suggested that neoadjuvant chemotherapy was more effective for advanced patients with histological grade 3 (7 –9). Higher mortality and local recurrence rates exist in patients with grade 3 STS. The proliferating cell nuclear antigen Ki-67 is currently one of the most reliable indicators of cell proliferation activity. It has been suggested that Ki-67 has better accuracy and reproducibility than the FNCLCC system in assessing tumor prognosis (10). Several long-term follow-up studies have shown that high Ki-67 expression in STS may be associated with higher mortality and recurrence rates (11,12). Thus, preoperative assessment of histologic grading and Ki-67 expression is essential for establishing individualized treatment plans and predicting prognosis in STS.

At present, preoperative evaluation of histological grade and Ki-67 expression mainly depends on puncture biopsy, but the limitations of the puncture site and the high heterogeneity of STS may lead to underestimate for the histological features (13). Diffusion-weighted imaging (DWI) has been widely used in the field of skeletal and muscular system lesions in recent years (14). DWI reflects the degree of diffusion of free water in the tissue and supplements functional information that is difficult to provide with conventional magnetic resonance imaging (MRI). The apparent diffusion coefficient (ADC) is used to quantitatively analyze the tissue diffusion information (15). In the identification of benign and malignant soft tissue tumors, histological grade, and evaluation of cell proliferation, a simple single index model has been proved to be reliable (16 –18). However, traditional ADC is affected by microcirculation perfusion, and this is calculated on the assumption that the water molecule is Gaussian, while the real diffusion of water molecules is irregular non-Gaussian motion. Intravoxel incoherent motion (IVIM) can separate perfusion and diffusion information in the tissue through multi-b-value imaging (19), and diffusion kurtosis imaging (DKI) can evaluate the most realistic non-Gaussian motion patterns of water molecules in tissues (20). These two imaging modes made up for the shortcomings of DWI from different perspectives. It has been shown that IVIM and DKI values reflect different histopathological features, and there are already large meta-analyses and multicenter studies about different histopathological parameters such as cell count and/or Ki-67 expression (21 –26).

Therefore, the aim of the present study was to evaluate the association between IVIM and DKI quantitative parameters and histopathological features in STS. We hypothesized that IVIM and DKI can provide more functional information than conventional MRI and traditional DWI to evaluate the histological grade and Ki-67 expression in STS.

Material and Methods

This retrospective case-control study was approved by the ethics committee. The need to obtain written informed consent from the participants was waived.

Study population

From September 2018 to March 2020, a total of 120 patients with suspected soft tissue masses were accepted in our hospital. In total, 90 patients underwent 3-T MRI including complete conventional MRI sequences, IVIM, and DKI images. The following exclusion criteria were used: (i) IVIM and DKI image artifacts were too hard to complete postprocessing analysis; (ii) patients who had received any relevant treatment before surgery; (iii) tumors occurring in the abdomen and pelvis; and (iv) patients whose final classification could not be determined after repeated observations by pathologists. Finally, 40 patients (22 men [mean age = 56.5 years, age range = 22–73 years], 18 women [mean age = 52 years, age range = 37–62 years]) were included in the study.

MRI acquisition

All MRI examinations were used through the GE Discovery 750W 3.0T MRI scanner (GE Healthcare, Milwaukee, USA). The 16-channel phased array abdomen and large flexible coil were selected. Conventional MRI scans used fast spin-echo (FSE) sequences and fat-suppressed (FS) sequences. The echo-planar imaging sequence was used for IVIM and DKI. The scanning layer thickness, layer spacing, layer number, and field of view were consistent with the cross-section T2-weighted (T2W) imaging sequence. The IVIM original image was collected using 8 b-values (0, 25, 50, 75, 100, 200, 500, and 800 s/mm²). The collection time was 2 min 33 s. Three b-values (0, 1000, and 2000 s/mm²) were used to collect DKI original images. Each b-value was applied with diffusion-sensitive gradient fields in 15 directions. The total acquisition time was 8 min 50 s. Finally, contrast-enhanced T1-weighted (T1W) imaging sequences were collected by injection of contrast agent through the cubital vein. Detailed MRI parameters are shown in Table 1.

Table 1.

MRI parameters for the different sequences.

Parameter	T1W	T2W	IVIM	DKI
TR (ms)	478–836	3042–5521	3000	5000
TE (ms)	14–18	70–82	70	80
NEX	1	1	2–4	4
Slice thickness (mm)	3–7	3–7	3–7	3–7
Matrix size	512 × 512	512 × 512	128 × 128	128 × 128
FOV (mm)	80–400	80–400	80–400	80–400

DKI, diffusion kurtosis imaging; FOV, field of view; IVIM, intravoxel incoherent motion; NEX, number of excitations; T1WI, T1-weighted imaging; T2WI, T2-weighted imaging; TE, echo time; TR, repetition time.

MRI analysis

Analysis of all MRI scans was independently performed by two skeletal muscle radiologists with 8 and 20 years of experience who were blinded to patients’ pathology results. All MRI scans were analyzed before surgery. The intra-group correlation coefficient (ICC) was used to assess the consistency of the two observers’ analysis of conventional MRI signs and the measurement of IVIM and DKI parameters (ICC < 0.40 = poor agreement; 0.41–0.59 = fair agreement; 0.60–0.74 = good agreement; 0.75–1.00 = excellent agreement).

IVIM and DKI images

The image postprocessing operation was performed by ADW4.7 workstation Functool software, which was provided by the GE original company.

IVIM: Multi-b-value images were postprocessed using Functool software. After adjusting the threshold and reducing the background noise, the single exponential model and the double exponential model were selected for image fitting. After manually delineating the region of interest (ROI), standard ADC, true diffusion (D), pseudo-diffusion (D*), and perfusion fraction (f) maps were obtained.

DKI: The kurtosis model was selected to fit multiple b-value (0, 1000, 2000) images, and 3D synchroview (GE Healthcare, Chicago, IL, USA) was used to ensure that the size and location of the ROI were completely consistent with the IVIM. Subsequently, mean kurtosis (MK) and mean diffusion (MD) maps were acquired.

ROI sketching principles: Standard ADC maps were selected for ROI positioning. The solid tumor region with the lowest ADC value (darkest) was chosen as the ROI, which corresponds to the highest DWI image signal region or the most obvious T1W imaging enhancement region. If there was no ADC low signal area, the largest section of the tumor was selected for delineation. Mucus, cystic changes, necrosis, calcification, and bleeding areas were avoided during the delineation process according to conventional MR sequences and enhancement sequence. It was essential to avoid the edges of the tumor to reduce the effects of partial volume effects.

Conventional MRI analysis

The tumor size, location, tumor heterogeneity, and tumor margins were recorded by two observers on conventional MRI. The maximum transverse diameter of the tumor was distinguished by 5 cm. The location of the tumor included subcutaneous and deep tissues. High T2W imaging heterogeneity is defined as more than half of tumors presenting mixed signals. Peritumoral enhancement, necrosis, and T2W imaging peritumoral edema were divided into present or absent.

Pathological analysis

Histopathological types, histological grade, and Ki-67 expression diagnosis of all tumors were jointly performed by two pathologists with >10 years of diagnostic experience. They were blinded to the patients’ clinical and MRI results. The degree of differentiation of tumor cells, the number of mitotic figures, and the area of tumor necrosis visible under the microscope were strictly determined according to the scoring standards of the FNCLCC system. All patients were divided into grade 1, grade 2, or grade 3. Grades 1 and 2 were classified as the low-grade group and grade 3 as the high-grade group. Immunohistochemical analysis was performed using Ki-67 antibody (1:200, Dako). A minimum of 1000 cells from five random high magnification fields were counted to identify Ki-67 positive cells present at a relatively uniform density. Their percentage of the total number of cells was calculated as Ki-67 positive expression.

Statistical analysis

Statistical analyses were performed using SPSS version 26.0 (IBM Corp., Chicago, IL, USA). The normality of all continuous variable data was tested. All MRI features in the high- and low-grade groups were compared. Categorical variables were analyzed using a chi-square test. For continuous variables, an independent sample t-test and Mann–Whitney U test were used. Pearson and Spearman analyses were performed on the correlation between the variables. The receiver operating characteristic (ROC) curve analysis was performed on the statistically significant parameters, and cutoff value, area under the curve (AUC), sensitivity, and specificity were used to evaluate the diagnostic efficacy of predicting the high-grade group.

Results

Fig. 1 shows the flow of patients included in this study. A total of 40 patients were selected for the study. There were 17 patients in the low-grade group (8 patients for grade 1 and 9 patients for grade 2) and 23 patients in the high-grade group (grade 3). Among the 40 lesions, Ki-67 expression was in the range of 10%–80%. The pathological results and locations of all lesions are shown in Table 2.

Fig. 1.

Flow chart.

Table 2.

Soft tissue sarcoma types, histology grades, and locations.

No.	Histopathological types	Histological grade	Ki-67 expression (%)	Location
1	Malignant peripheral nerve sheath tumor	Grade 1	10	R thigh
2	Malignant peripheral nerve sheath tumor	Grade 1	15	R shoulder
3	Adult fibrosarcoma	Grade 1	25	L hip
4	Adult fibrosarcoma	Grade 1	25	L thigh
5	Myxofibrosarcoma	Grade 1	30	R knee
6	Low-grade fibromyxoid sarcoma	Grade 1	30	L thigh
7	Myxoid liposarcoma	Grade 1	50	L knee
8	Myxoid liposarcoma	Grade 1	60	L thigh
9	Myxoid liposarcoma	Grade 2	15	R thigh
10	Myxoid liposarcoma	Grade 2	20	L upper arm
11	Myxofibrosarcoma	Grade 2	30	L forearm
12	Adult fibrosarcoma	Grade 2	35	R thigh
13	Synovial sarcoma	Grade 2	35	R ankle
14	Synovial sarcoma	Grade 2	40	L foot
15	Malignant peripheral nerve sheath tumor	Grade 2	45	R thigh
16	Leiomyosarcoma	Grade 2	50	R thigh
17	Malignant giant cell tumor of tendon sheath	Grade 2	55	L thigh
18	Pleomorphic liposarcoma	Grade 3	15	L foot
19	Pleomorphic liposarcoma	Grade 3	20	R lower leg
20	Pleomorphic liposarcoma	Grade 3	25	L thigh
21	Pleomorphic liposarcoma	Grade 3	25	L upper arm
22	Dedifferentiated liposarcoma	Grade 3	30	R thigh
23	Dedifferentiated liposarcoma	Grade 3	30	R thigh
24	Dedifferentiated liposarcoma	Grade 3	35	L lower leg
25	Undifferentiated pleomorphic sarcoma	Grade 3	35	L thigh
26	Undifferentiated pleomorphic sarcoma	Grade 3	40	L ankle
27	Undifferentiated pleomorphic sarcoma	Grade 3	40	R shoulder
28	Undifferentiated pleomorphic sarcoma	Grade 3	40	R thigh
29	Undifferentiated pleomorphic sarcoma	Grade 3	45	L knee
30	Undifferentiated pleomorphic sarcoma	Grade 3	45	R lower leg
31	Myxofibrosarcoma	Grade 3	50	L shoulder
32	Myxofibrosarcoma	Grade 3	55	R lower leg
33	Adult fibrosarcoma	Grade 3	55	R lower leg
34	Leiomyosarcoma	Grade 3	55	L hip
35	Leiomyosarcoma	Grade 3	60	R thigh
36	Leiomyosarcoma	Grade 3	65	R knee
37	Malignant peripheral nerve sheath tumor	Grade 3	70	L thigh
38	Synovial sarcoma	Grade 3	70	L thigh
39	Synovial sarcoma	Grade 3	75	L upper arm
40	Synovial sarcoma	Grade 3	80	R thigh

Spearman correlation coefficient showed that D value was inversely correlated with Ki-67 expression (r = –0.372; P = 0.018). MK value showed a moderate positive correlation with Ki-67 expression (r = 0.570; P < 0.001). However, standard ADC, D*, f, and MD values did not show any significant correlations (P > 0.05). Furthermore, Pearson examination revealed a moderate positive correlation between Ki-67 expression and histological grading (r = 0.364; P = 0.021).

D and MD values showed a negative correlation with histological grade (r = –0.351; r = –0.422). MK value showed a positive correlation with histological grade (r = 0.454; P = 0.003). Standard ADC, D*, and f values did not show any significant correlations (P > 0.05).

Among all conventional MRI features, only peritumoral enhancement was statistically different between the high- and low-grade groups (P = 0.024). The IVIM and DKI parameters measured by two observers showed excellent consistency (ICC > 0.940).

As shown in Table 3, the high-grade group had a significantly lower D value (P = 0.014) than the low-grade group; however, standard ADC, D*, and f values had no significant difference between the high- and low-grade groups. When the cutoff value was 1.306 × 10⁻³ mm²/s, the AUC, sensitivity, and specificity for differentiating between high- and low-grade STS were 0.729, 73.9%, and 70.6%, respectively (Figs. 2 and 3).

Fig. 2.

A myxoid liposarcoma in a 35-year-old patient. (a)The tumor was located in deep tissue of the thigh (red arrow). It performed slightly heterogeneous signal intensities (SI) at T2-weighted imaging. (b) After injection of contrast agent, the lesion was evenly strengthened, and no peritumoral enhancement was found. (c, d) IVIM (b-value = 500 s/mm²) and DKI (b-value = 2000 s/mm²) images identified high signal intensities. (e–g) The ROI was manually placed at the lowest signal of the ADC map. The D, MK, and MD maps and corresponding values are as follows:1.53 × 10⁻³mm², 0.47, and 1.61 × 10⁻³mm², respectively. (h) Postoperative pathological examination (hematoxylin & eosin–stained, ×400) confirmed the lesion as low-grade (grade 2) soft tissue sarcoma. ADC, apparent diffusion coefficient; DKI, diffusion kurtosis imaging; IVIM, intravoxel incoherent motion; MK, mean diffusion; MK, mean kurtosis; ROI, region of interest.

Fig. 3.

A leiomyosarcoma in a 55-year-old patient. (a) The tumor was located in the deep tissue of the thigh (red arrow). It performed highly heterogeneous signal intensities (SI) at T2-weighted imaging. (b) After injection of contrast agent, the lesion was performed unevenly strengthened and peritumoral enhancement was demonstrated on T1-weighted imaging (black arrow). (c, d) IVIM (b = 500 s/mm²) and DKI (b-value = 2000 s/mm²) images identified high signal intensities. (e–g) The ROI was manually placed at the lowest signal of the ADC map. The D, MK, and MD maps and corresponding values are as follows:1.12 × 10⁻³mm², 0.68, 1.39 × 10⁻³mm², respectively. (h) Postoperative pathological examination (hematoxylin & eosin–stained, ×400) confirmed the lesion as high-grade (grade 3) soft tissue sarcoma. ADC, apparent diffusion coefficient; DKI, diffusion kurtosis imaging; IVIM, intravoxel incoherent motion; MK, mean diffusion; MK, mean kurtosis; ROI, region of interest.

Table 3.

Quantitative and qualitative MRI characteristics of soft tissue sarcomas.

MRI parameters	Low-grade group	High-grade group	P
Standard ADC	1.51 ± 0.07	1.48 ± 0.05	0.174
D	1.36 ± 0.12	1.26 ± 0.13	0.014*
D*	22.83 ± 13.42	28.27 ± 11.21	0.171
F	14.88 ± 3.31	16.86 ± 5.13	0.173
MK	0.51 ± 0.08	0.63 ± 0.11	<0.001*
MD	1.53 ± 0.10	1.46 ± 0.08	0.019*
Tumor size (cm)			0.107
<5	9/17 (53)	6/23 (26)
≥5	8/17 (47)	17/12 (74)
Location			0.216
Subcutaneous	11/17 (64)	10/23 (43)
Deep	6/17 (36)	13/23 (57)
Peritumoral enhancement			0.024*
Absent	12/17 (71)	7/23 (30)
Present	5/17 (29)	16/23 (70)
Necrosis			0.198
Absent	8/17 (47)	6/23 (26)
Present	9/17 (53)	17/23 (74)
T2WI heterogeneity			0.202
Highly heterogeneous	5/17 (29)	12/23 (52)
Slightly heterogeneous	12/17 (71)	11/23 (48)
T2WI peritumoral edema			0.337
Absent	9/17 (53)	8/23 (35)
Present	8/17 (47)	15/23 (65)

Values are given as n (%) or mean ± SD.

*Statistically significant at P < 0.05.

ADC, apparent diffusion coefficient; MD, mean diffusion; MK, mean kurtosis; MRI, magnetic resonance imaging; T1WI, T1-weighted imaging; T2WI, T2-weighted imaging.

The high-grade group had a significantly higher MK value (P < 0.001) and lower MD value (P = 0.019) than the low-grade group (Table 2). The MK value showed the best diagnostic performance. The AUC of the MK value for differentiating between high- and low-grade group was 0.812. With a cutoff value of 0.552, the sensitivity and specificity were 78.3% and 76.5%, respectively. The AUC of the MD value for differentiating between the high- and low-grade group was 0.694. At a cutoff value of 1.518 × 10⁻³ mm²/s, the sensitivity and specificity were 78.3% and 64.7%, respectively (Figs. 2 and 3).

Table 4 showed the value of joint diagnostics with statistically significant parameters. The combination of MK and MD yielded the highest specificity (88.2%). The combination of D, MK, and MD yielded the best AUC value (0.900) and sensitivity (95.7%) than other models.

Table 4.

Diagnostic performance of IVIM and DKI parameters for differentiating between high- and low-grade groups.

Parameter(s)	AUC	Specificity	Sensitivity	Accuracy	PPV	NPV
D	0.729	12/17 (70.6)	17/23 (73.9)	29/40 (72.5)	17/22 (77.3)	12/18 (66.7)
MK	0.812	13/17 (76.5)	18/23 (78.3)	31/40 (77.5)	18/22 (81.8)	13/18 (72.2)
MD	0.694	11/17 (64.7)	18/23 (78.3)	29/40 (72.5)	18/24 (75.0)	11/16 (68.8)
D + MK	0.862	13/17 (76.5)	19/23 (82.6)	32/40 (80.0)	19/23 (82.6)	13/17 (76.5)
D + MD	0.806	12/17 (70.6)	19/23 (82.6)	31/40 (77.5)	19/24 (79.2)	12/16 (75.0)
MK + MD	0.872	15/17 (88.2)	20/23 (87.0)	35/40 (87.0)	20/22 (90.9)	15/18 (83.3)
D + MK + MD	0.900	14/17 (82.4)	22/23 (95.7)	36/40 (90.0)	22/25 (88.0)	14/15 (93.3)

Values are given as n (%).

AUC, area under the curve; DKI, diffusion kurtosis imaging; IVIM, intravoxel incoherent motion; MD, mean diffusion; MK, mean kurtosis; NPV, negative predictive value; PPV, positive predictive value.

Discussion

It has become a hot research topic in recent years to evaluate more accurately the histological grading and cell proliferation ability of STS by MRI. With regard to the histological grading, we found that conventional MRI has potential markers suggesting high-grade lesions and D values obtained from IVIM, and MK and MD values obtained from DKI can be used as reliable predictors of grade 3 STS. With regard to the proliferative capacity of sarcoma cells, D and MK values proved to be very useful in predicting Ki-67 expression.

IVIM is an optimized sequence model based on traditional DWI. In the present study, we found that the standard ADC value obtained from the traditional single-index model cannot predict grade 3 STS, which is inconsistent with the research by Chhabra et al. (18). We speculated that the diversity of histological types and the small sample size might have contributed to this result. In addition, the measurement of standard ADC could be affected by many factors, such as the steroid component in the mucus matrix, and the rich capillaries might cause inaccuracies in the ADC value. The D value could reflect the diffusion information of pure water molecules. Our results revealed that the D value could predict grade 3 STS, which is consistent with the results of Togao et al. (27) and Zhu et al. (28). We argued that grade 3 STS was often accompanied by more vigorous mitosis, which led to increases in the number of tumors and the size of the nucleus. The narrowing gap between the inside and outside of the cell further limits the diffusion movement of water molecules. Therefore, the grade 3 STS has a lower D value than grade 2 and grade 3 STS.

A negative correlation between ADC values and Ki-67 expression has been demonstrated in solid tumors such as breast cancer, hepatocellular carcinoma, and cervical cancer (29 –32). With regard to STS, Lee et al. first showed a weak negative correlation between ADC values and Ki-67 (17), while Yuan et al. explored a moderate negative correlation between ADC and D values and Ki-67 expression in a mouse model of rhabdomyosarcoma (33). Our study showed that ADC values did not correlate with Ki-67 expression, while D values decreased as Ki-67 expression increased. Although it is theoretically suggested that increased cell number and cell density per unit volume leads to a narrowing of the cell gap and thus causes a restriction in the true diffusion of water molecules (34), the instability brought about by the histological type cannot be ignored. Targeted studies for each subtype of sarcoma may be more clinically instructive.

The D* and f values reflect the microcirculation perfusion in the tissue, which is mainly affected by the number and length of capillaries and the permeability of the blood vessels (35). Our research demonstrated that both were of no value in distinguishing histological grade and Ki-67 expression, which was consistent with a previous study conducted on hepatocellular carcinoma (36). We reviewed several studies that applied IVIM to soft tissue tumors and found that scholars are controversial about the ability of D* and f to reflect the microcirculation of soft tissue tumors. According to the research by Wu et al. (37) and Du et al. (38), D* and f values could distinguish the blood supply status of benign and malignant soft tissue tumors. Conversely, Lim et al. got completely different results (39). We believed that the setting of b-values might cause this difference. Wu selected 9 b-values <200 s/mm² to complete IVIM image acquisition and obtained stable D* and f value measurement results. However, only two b-values <100–200 s/mm² were selected in the study by Lim et al., and the D* and f values obtained were unstable. In our study, 5 b-values were selected for the measurement of perfusion parameters, which may be the critical point for modeling stability. The larger standard deviation in the final measurement results also implies that the accurate modeling of the D* value requires more b-values to guarantee. In addition, the research by Yuan et al. indicated that D* could reflect the expression of VEGF in a mouse model of rhabdomyosarcoma, while the f value could not (40). Marzi et al. argued that there was no significant correlation between the perfusion parameters derived from IVIM and DCE-MRI parameters in soft tissue tumors (41). Therefore, we believe that the reliability and stability of the D* and f values used to reflect the perfusion level of STS need to be further confirmed by the selection of more b-values, accurate modeling, and narrowing the scope of the study.

Our research revealed that both MK and MD values show excellent value in predicting STS at grade 3; MK showed the highest diagnostic accuracy. Our findings are consistent with the study by Falk et al. (42). According to the research by Crombe et al., tumor heterogeneity is an independent predictor of grade 3 STS (43), and we tried to quantitatively evaluate this phenomenon using MK values. At the same time, our research shows that the diagnostic ability of the MK value is much higher than the MD value. This may mean that the histological grade of STS is not only reflected in cell density. Grade 3 STS can cause changes in tissue hypoxia, abnormal angiogenesis, necrosis, bleeding, and so on while growing rapidly, and MK can more accurately reflect the complexity of the tumor. The MD value measurement is more susceptible to interference, such as mucus and perfusion volume. Therefore, we believe that the MK value is more advantageous in evaluating high-grade STS.

The value of kurtosis parameters in assessing cell proliferation capacity in epithelial ovarian tumors and breast cancer has been investigated (44,45). To our knowledge, this is the first time that kurtosis parameters have been used to explore cell proliferative capacity in STS. We found that MK values showed the best correlation out of Ki-67 expression. Specifically, MK values increased with significantly higher Ki-67 expression, which is consistent with previous findings (44). DKI has the advantage of reflecting the anisotropy of the diffusive movement of water molecules within the tissue, whereas conventional DWI and IVIM imaging assess molecular movement within the tissue by isotropic techniques (46). However, the use of DKI in soft tissue tumors is still relatively rare, and the optimal scanning protocol for acquiring high b-value images is still in the developmental stage (47,48). Nevertheless, in this study, DKI remains the best imaging tool to evaluate the histological grading and cell proliferation capacity of STS.

Through the multi-parameter joint analysis, the predictive ability of quantitative parameters for grade 3 STS has been further improved. Diagnostic sensitivity increased from 78.26% to 95.65%, and diagnostic specificity was raised from 76.47% to 88.24%. We found that all parameters overlapped between the two groups of STS, and a multi-parameter combination can effectively avoid this interference.

In previous studies, peritumoral enhancement is considered a reliable predictor of grade 3 STS, which is consistent with our findings (43,49). However, other conventional MRI features, such as T2W imaging heterogeneity and necrosis, were unable to predict grade 3 STS in this study. We infer that the sample size may have a certain impact. At the same time, this may also indicate that the stability of qualitative diagnosis is not as good as quantitative analysis.

The present study has some limitations. First, the sample size was small. The variety of STS categories might affect our results. Second, the IVIM scanning scheme used in this study may not be the best choice, and we think this will be a hot issue in the next step. Finally, we did not separately analyze the STS rich in mucus components. The steroid components may affect the accuracy of the ADC results.

In conclusion, our study confirms the reliability of IVIM and DKI in predicting histological grade and Ki-67 expression; the quantitative parameters obtained from IVIM and DKI have high stability. In addition, the DKI model is more accurate than the IVIM model. The MK value had the best correlation with histological grading and cell proliferation capacity. Future research is needed to optimize MRI protocols to provide an imaging basis for clinical decision-making.

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was supported in part by grants from the National Natural Science Foundation of China (#81771804).

ORCID iDs

Yifeng Zhu

Shaowu Wang

References

Carbonnaux

Brahmi

Schiffler

, et al. Very long-term survivors among patients with metastatic soft tissue sarcoma. Cancer Med 2019;8:1368–1378.

Coindre

Terrier

Guillou

, et al. Predictive value of grade for metastasis development in the main histologic types of adult soft tissue sarcomas: a study of 1240 patients from the French Federation of Cancer Centers Sarcoma Group. Cancer 2001;91:1914–1926.

Siddiqi

Kim

Jede

, et al. Association of core needle biopsy tract resection with local recurrence in extremity soft tissue sarcoma. Skeletal Radiol 2017;46:507–512.

Guillou

Coindre

Bonichon

, et al. Comparative study of the National Cancer Institute and French Federation of Cancer Centers Sarcoma Group grading systems in a population of 410 adult patients with soft tissue sarcoma. J Clin Oncol 1997;15:350–362.

Sundby Hall

Bruland

Bjerkehagen

, et al. Adjuvant chemotherapy and postoperative radiotherapy in high-risk soft tissue sarcoma patients defined by biological risk factors-A Scandinavian Sarcoma Group study (SSG XX). Eur J Cancer 2018;99:78–85.

Reynoso

Subbiah

Trent

, et al. Neoadjuvant treatment of soft-tissue sarcoma: a multimodality approach. J Surg Oncol 2010;101:327–333.

Gronchi

Ferrari

Quagliuolo

, et al. Histotype-tailored neoadjuvant chemotherapy versus standard chemotherapy in patients with high-risk soft-tissue sarcomas (ISG-STS 1001): an international, open-label, randomised, controlled, phase 3, multicentre trial. Lancet Oncol 2017;18:812–822.

Pasquali

Gronchi

. Neoadjuvant chemotherapy in soft tissue sarcomas: latest evidence and clinical implications. Ther Adv Med Oncol 2017;9:415–429.

Saponara

Stacchiotti

Casali

, et al. (Neo)adjuvant treatment in localised soft tissue sarcoma: the unsolved affair. Eur J Cancer 2017;70:1–11.

10.

Tanaka

Hasegawa

Nojima

, et al. Prospective evaluation of Ki-67 system in histological grading of soft tissue sarcomas in the Japan Clinical Oncology Group Study JCOG0304. World J Surg Oncol 2016;14:110.

11.

Hasegawa

Yamamoto

Yokoyama

, et al. Prognostic significance of grading and staging systems using MIB-1 score in adult patients with soft tissue sarcoma of the extremities and trunk. Cancer 2002;95:843–851.

12.

Hoos

Stojadinovic

Mastorides

, et al. High Ki-67 proliferative index predicts disease specific survival in patients with high-risk soft tissue sarcomas. Cancer 2001;92:869–874.

13.

Yang

Frassica

Fayad

, et al. Analysis of nondiagnostic results after image-guided needle biopsies of musculoskeletal lesions. Clin Orthop Relat Res 2010;468:3103–3111.

14.

Oka

Yakushiji

Sato

, et al. Ability of diffusion-weighted imaging for the differential diagnosis between chronic expanding hematomas and malignant soft tissue tumors. J Magn Reson Imaging 2008;28:1195–1200.

15.

Subhawong

Jacobs

Fayad

. Diffusion-weighted MR imaging for characterizing musculoskeletal lesions. Radiographics 2014;34:1163–1177.

16.

Choi

Lee

Song

, et al. Diagnostic performance of diffusion-weighted (DWI) and dynamic contrast-enhanced (DCE) MRI for the differentiation of benign from malignant soft-tissue tumors. J Magn Reson Imaging 2019;50:798–809.

17.

Lee

Yoon

Seo

, et al. Soft tissue sarcoma: DWI and DCE-MRI parameters correlate with Ki-67 labeling index. Eur Radiol 2020;30:914–924.

18.

Chhabra

Ashikyan

Slepicka

, et al. Conventional MR and diffusion-weighted imaging of musculoskeletal soft tissue malignancy: correlation with histologic grading. Eur Radiol 2019;29:4485–4494.

19.

Le Bihan

Breton

Lallemand

, et al. Separation of diffusion and perfusion in intravoxel incoherent motion MR imaging. Radiology 1988;168:497–505.

20.

Raab

Hattingen

Franz

, et al. Cerebral gliomas: diffusional kurtosis imaging analysis of microstructural differences. Radiology 2010;254:876–881.

21.

Zhang

Chen

, et al. Grading and proliferation assessment of diffuse astrocytic tumors with monoexponential, biexponential, and stretched-exponential diffusion-weighted imaging and diffusion kurtosis imaging. Eur J Radiol 2018;109:188–195.

22.

Woo

Lee

Yoon

, et al. Intravoxel incoherent motion diffusion-weighted MR imaging of hepatocellular carcinoma: correlation with enhancement degree and histologic grade. Radiology 2014;270:758–767.

23.

Wang

, et al. Quantitative evaluation of intravoxel incoherent motion and diffusion Kurtosis imaging in assessment of pathological grade of clear cell renal cell carcinoma. Acad Radiol 2019;27:7.

24.

Xiao

Tang

Zheng

, et al. Diffusion kurtosis imaging and intravoxel incoherent motion in differentiating nasal malignancies. Laryngoscope 2020;130:E727–E735.

25.

Xiao

Tang

Zhang

, et al. Whole-tumor histogram analysis of monoexponential and advanced diffusion-weighted imaging for sinonasal malignant tumors: correlations with histopathologic features. J Magn Reson Imaging 2020;51:273–285.

26.

Zhang

Yang

, et al. Diffusion metrics for staging pancreatic fibrosis and correlating with epithelial-mesenchymal transition markers in a chronic pancreatitis rat model at 11.7 T MRI. J Magn Reson Imaging 2020;52:197–206.

27.

Togao

Hiwatashi

Yamashita

, et al. Differentiation of high-grade and low-grade diffuse gliomas by intravoxel incoherent motion MR imaging. Neuro Oncol 2016;18:132–141.

28.

Zhu

, et al. Value of intravoxel incoherent motion in assessment of pathological grade of clear cell renal cell carcinoma. Acta Radiol 2018;59:121–127.

29.

Catalano

Horn

Signore

, et al. PET/MR in invasive ductal breast cancer: correlation between imaging markers and histological phenotype. Br J Cancer 2017;116:893–902.

30.

Mori

Ota

Mugikura

, et al. Luminal-type breast cancer: correlation of apparent diffusion coefficients with the Ki-67 labeling index. Radiology 2015;274:66–73.

31.

Huang

Meng

, et al. Correlations between ADC values and molecular markers of Ki-67 and HIF-1alpha in hepatocellular carcinoma. Eur J Radiol 2015;84:2464–2469.

32.

, et al. Intravoxel incoherent motion combined with dynamic contrast-enhanced perfusion MRI of early cervical carcinoma: correlations between multimodal parameters and HIF-1alpha expression. J Magn Reson Imaging 2019;50:918–929.

33.

Yuan

Zeng

Liu

, et al. DWI and IVIM are predictors of Ki67 proliferation index: direct comparison of MRI images and pathological slices in a murine model of rhabdomyosarcoma. Eur Radiol 2020;30:1334–1341.

34.

Surov

Clauser

Chang

, et al. Can diffusion-weighted imaging predict tumor grade and expression of Ki-67 in breast cancer? A multicenter analysis. Breast Cancer Res 2018;20:58.

35.

Cho

Kim

Jensen

, et al. A versatile flow phantom for intravoxel incoherent motion MRI. Magn Reson Med 2012;67:1710–1720.

36.

Zhu

Liu

Wei

, et al. Intravoxel incoherent motion diffusion-weighted magnetic resonance imaging for predicting histological grade of hepatocellular carcinoma: comparison with conventional diffusion-weighted imaging. World J Gastroenterol 2018;24:929–940.

37.

Zhang

Liang

, et al. Intravoxel incoherent motion MRI for the differentiation of benign, intermediate, and malignant solid soft-tissue tumors. J Magn Reson Imaging 2017;46:1611–1618.

38.

Zhang

, et al. Intravoxel incoherent motion MR imaging: comparison of diffusion and perfusion characteristics for differential diagnosis of soft tissue tumors. Medicine (Baltimore 2015;94:e1028.

39.

Lim

Jee

Jung

, et al. Intravoxel incoherent motion diffusion-weighted MR imaging for differentiation of benign and malignant musculoskeletal tumours at 3T. Br J Radiol 2018;91:20170636.

40.

Yuan

Zeng

Zhang

, et al. Intravoxel incoherent motion diffusion-weighted imaging assessment of microvascular characteristics in the murine embryonal rhabdomyosarcoma model. Acta Radiol 2020;61:260–266.

41.

Marzi

Stefanetti

Sperati

, et al. Relationship between diffusion parameters derived from intravoxel incoherent motion MRI and perfusion measured by dynamic contrast-enhanced MRI of soft tissue tumors. NMR Biomed 2016;29:6–14.

42.

Falk Delgado

Nilsson

van Westen

, et al. Glioma grade discrimination with MR diffusion Kurtosis imaging: a meta-analysis of diagnostic accuracy. Radiology 2018;287:119–127.

43.

Crombe

Marcellin

Buy

, et al. Soft-tissue sarcomas: assessment of MRI features correlating with histologic grade and patient outcome. Radiology 2019;291:710–721.

44.

Huang

Lin

, et al. Diffusion kurtosis at 3.0T as an in vivo imaging marker for breast cancer characterization: correlation with prognostic factors. J Magn Reson Imaging 2019;49:845–856.

45.

Zhao

Qiang

, et al. Diffusion kurtosis imaging for differentiating borderline from malignant epithelial ovarian tumors: a correlation with Ki-67 expression. J Magn Reson Imaging 2017;46:1499–1506.

46.

Bai

Lin

Tian

, et al. Grading of gliomas by using monoexponential, biexponential, and stretched exponential diffusion-weighted MR imaging and diffusion Kurtosis MR imaging. Radiology 2016;278:496–504.

47.

Yang

Wang

, et al. Soft tissue sarcomas: IVIM and DKI correlate with the expression of HIF-1alpha on direct comparison of MRI and pathological slices. Eur Radiol 2021;31:4669–4679.

48.

Liu

Xiong

, et al. Intravoxel incoherent motion and diffusion kurtosis imaging for discriminating soft tissue sarcoma from vascular anomalies. Medicine (Baltimore) 2018;97:e13641.

49.

Zhao

Ahlawat

Farahani

, et al.

Can MR imaging be used to predict tumor grade in soft-tissue sarcoma?

Radiology 2014;272:192–201.