Introduction

Rectal cancers affect the rectum, and are generally adenocarcinomas (1). Nearly 70% of patients are >65 years old, and rectal cancer occurs rarely in patients who are under 40 years of age; in addition, men are predominantly affected (2). Probable risk factors include age, male gender, colon polyps, a history of colorectal cancer, a history of inflammatory bowel disease, hereditary syndromes, lifestyle factors (diet, alcohol, obesity, sedentary lifestyle and smoking), and a history of diabetes mellitus (1). Rectal cancer is usually managed using a combination of surgery, chemotherapy, targeted therapy and radiation therapy (3). The five-year survival rates following surgical resection are 85–95% for stage I, 60–80% for stage II, and 30–60% for stage III cancer. With the aging of the Chinese society and a Westernization of the diet, the incidence of rectal cancer in China is increasing (4). Therefore, finding novel means to efficiently detect and diagnose rectal cancer are required.

Dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) is a relatively novel imaging modality that demonstrates the capillary blood flows (5,6). In DCE-MRI, the distribution of the contrast agent is repeatedly evaluated, allowing the evaluation of the tumor microcirculation in vivo and enabling the malignancy or benignancy of the tumor to be quantitatively distinguished (6,7). A number of previous studies have shown the diagnostic value of DCE-MRI for prostate cancer (8–10), for the evaluation of pancreatic cancer (11), and for evaluating the efficacy of cancer drugs (12–14). Only one previous study has used DCE-MRI to evaluate the tumor drug response in colorectal cancer (15), and, at present, no published studies have been concerned with the diagnostic value of DCE-MRI for rectal cancer.

Therefore, the two-fold aim of the present study was to use DCE-MRI to compare the tumor perfusion parameters with postoperative pathology results in order to elucidate the characteristics of the DCE-MRI parameters in rectal cancer, and to explore the application of DCE-MRI in the preoperative diagnosis of rectal cancer.

Materials and methods

Subjects

Patients with a suspected rectal lesion (blood in stool, changes of bowel habits, and/or suspicious masses identified by finger examination or colonoscopy) were recruited at the Anorectal Surgery Department, Changhai Hospital Affiliated to the Second Military Medical University (Shanghai, China), between December 2013 and February 2015. All subjects underwent DCE-MRI and postoperative pathology. Inclusion criteria were: i) a single lesion on conventional imagery; and ii) no chemotherapy or radiotherapy was administered prior to DCE-MRI. Patients were excluded if: i) DCE-MRI images were of poor quality; ii) images did not include the whole lesion; iii) artifacts affected measurements; or iv) lesions were too small to be accurately measured.

Finally, 55 patients were included in the study, 40 of whom had been diagnosed with pathology-confirmed rectal cancer, whereas the remaining 15 subjects were included as controls. These 15 controls included two cases of chronic inflammation, one case of rectum hemorrhoids, one case of isolated hamartoma, six cases of tubular adenoma, three cases of tubular villous adenoma and two healthy controls.

The present study was approved by the Committee on Ethics of Biomedical Research, Second Military Medical University (Shanghai, China). All subjects signed a written informed consent.

MRI scanning

MRI was performed using a Skyra 3.0 T MRI scanner (Siemens AG, Erlangen, Germany) with a pelvic phased array coil. All subjects had fasted for 4 h prior to scanning. Routine pelvic MRI and DCE-MRI sequences were used for scanning. The routine sequences included sagittal T2 weighted imaging (T2WI) with fat suppression, axial T1WI, T2WI, diffusion (D)WI, and coronal T2WI. DCE-MRI was used to scan the same axial layers as T2WI with the following scanning parameters: Repetition time (TR) 3.33 msec, echo time (TE) 1.23 msec, flip angle 9°, field of view (FOV) 36 cm, matrix 125×192, layer thickness 3 mm and 30 layers; parallel acquisition by CAIPIRINHA and acceleration factor R=2. Each scan lasted for 5 sec, and 75 scans were performed with free respiration. Following the initial scan, the contrast agent, gadolinium-diethylenetriaminepentacetate (Beijing Hokuriku Pharmaceutical Co., Ltd., China; volume of 15–20 ml, 0.2 mmol/kg) was injected intravenously at 3 ml/sec using high-pressure syringes. Subsequently, 20 ml saline solution of 0.90% (w/v) NaCl was injected at the same rate.

Measurements and data analysis

All data obtained from the DCE sequences were sent to the Tissue 4D post-processing workstation (Siemens AG) for post-processing. Parameter calculations were performed using the Tofts pharmacokinetic model (16,17). On the DCE images, an area resembling a circle, including all rectal and surrounding mesangial regions, was selected to generate DCE-MRI pseudo-color images. The images were examined in a blind manner by two radiologists specializing in MRI, with 5 years of experience. By comparing with the T2WI images, the radiologists selected the plane sections where the lesions were relatively large, based on the morphology, sites, infiltration depth, enhancement patterns and invasion into surrounding regions to manually place regions of interest (ROI), with areas no less than 1 cm2. The perfusion parameters (16,17) were measured from three plane sections, and the average was used for analysis: i) volume transfer constant from the plasma compartment to the extravascular extracellular space (Ktrans); ii) rate constant for transfer between the extravascular extracellular space and the blood compartment (Kep); iii) volume of extravascular extracellular space per unit volume of tissue (Ve); and iv) the initial area under the enhancement curve (iAUC). Time-signal intensity curves were plotted.

Patients with rectal cancer were divided into three groups, according to the tumor differentiation based on the postoperative pathology reports evaluated in a blind manner by two pathologists with 5 years of experience: i) highly differentiated; ii) moderately differentiated; and iii) poorly differentiated (18).

Statistical analysis

SPSS version 16.0 software (SPSS, Inc., Chicago, IL, USA) was used for the statistical analysis. Data are expressed as the means ± standard deviation. Due to the heterogeneity of variance in the parameters between patients with rectal cancer and controls, the Mann-Whitney U test was used to assess the differences between the two groups. Intraclass correlation coefficients (ICCs) were used to estimate the consistency of the results from the two radiologists. To determine the diagnostic power of Ktrans, Kep, Ve, and iAUC, the receiver operator characteristics (ROC) curve method was used, and areas under the curve AUCs were calculated. Optimal cut-off values were selected to maximize specificity and sensitivity. The H test was used to compare DCE-MRI parameters among different degrees of differentiation. The correlations between parameters and differentiation of rectal cancer were analyzed using the Spearman's rank correlation test. P<0.05 was considered to indicate a statistically significant difference.

Results

Characteristics of the patients

Among the 40 patients with rectal cancer, there were 20 males and 20 females, aged 57±9 (range 25–68) years. Among the 15 controls, there were eight males and seven females, aged 56±7 (range 44–70) years. No significant differences in gender or age were identified between the two groups.

DCE-MRI manifestation of rectal cancer

All 40 cases of rectal cancer had a single lesion, of which 26 cases exhibited space-occupying masses, 10 revealing an irregular thickening of the local intestinal wall, and four cases showing abnormal local nodular signals. All lesions demonstrated equal or lower T1WI signals, equal or higher T2WI signals, equal or higher fat-suppression sequence signals, and high DWI signals. Significantly enhanced lesions were capable of being viewed on DCE-MRI. The Ktrans, Kep, Ve, and iAUC values in rectal cancer lesions were all significantly higher compared with the controls (all P<0.05) (Table I). The pseudo-color images of each parameter are shown in Figs. 1 and 2. The time-signal intensity curves showed an outflow pattern: The signal peaked shortly following the injection of the contrast agent, and subsequently quickly decayed by >10% compared with the signal intensity in the mid- and late-stage of the enhancement after the contrast agent was injected (Fig. 1G). On the other hand, in the controls, the time-signal intensity curves revealed a comparatively smaller elevation of signal intensity at the early stage, and a plateau at the later stage (Fig. 2G).

Figure 1. MRI images from the same patient with rectal cancer, a female aged 48 years with blood in the stool for 6 months. (A) A T2WI cross section, revealing the local occupation of the rear intestinal wall. (B) DCE-MRI scanning image showing a marked enhancement of the rear intestinal wall, which was slightly higher than the surrounding regions. (C) DCE-MRI pseudo-color images of Ktrans. The yellowish-green area indicates the local occupation of the intestinal wall. Ktrans of the lesion area was 0.208, which was higher compared with the surrounding intestinal wall tissue (blue). (D) DCE-MRI pseudo-color images of Kep. The red and yellow areas indicate the lesion area, where Kep was 0.588. (E) DCE-MRI pseudo-color images of Ve. The yellowish-green area indicates the lesion area, where Ve was 0.370. (F) DCE-MRI pseudo-color images of iAUC. The red and yellow areas indicate the lesion area, where iAUC was 37.009, which was higher compared with the surrounding intestinal wall tissue (blue). (G) The DCE-MRI time-signal intensity curve of the patient, where the dashed line indicates the connection between data points and the solid line indicates the fitted curve, reveals an outflow pattern. (H) Results of the postoperative pathology test (hematoxylin and eosin staining, ×200). The tumor tissue showed a papillary and mesh-like alignment. Tumor cells were cubic and of a column-like shape, with big, atypical and deeply stained nuclei, which infiltrated into the superficial muscle layer of the intestinal wall. The patient was diagnosed with moderately differentiated rectal adenocarcinoma. DCE-MRI, dynamic contrast-enhanced magnetic resonance imaging; Ktrans, volume transfer constant from the plasma compartment to the extravascular extracellular space; Kep, rate constant for transfer between extravascular extracellular space and the blood compartment; Ve, volume of extravascular extracellular space per unit volume of tissue; iAUC, initial area under enhancement curve.

Figure 2. MRI images from the same patient in the control group: A male, aged 35 years, with a feeling of incomplete defecation for one year, occasional anal pain, and slight blood in the stool. Swollen masses were identified by finger examination. (A) A T2WI cross-section showing abnormal occupation signals on the right side of the intestinal wall. (B) DCE-MRI scanning images showing similar enhancement of the space occupation of the selected intestinal wall compared with the surrounding regions. (C) DCE-MRI pseudo-color images of Ktrans. The blue color was evenly distributed along the intestine, and the Ktrans of the ROI was 0.100. (D) DCE-MRI pseudo-color images of Kep. The lesion area was blue, where Kep was 0.428. (E) DCE-MRI pseudo-color images of Ve. The lesion area was blue, where Ve was 0.336. (F) DCE-MRI pseudo-color images of iAUC. The lesion area was blue, where iAUC was 18.834, which was comparable with the surrounding intestinal wall tissue. (G) The DCE-MRI time-signal intensity curve of the control revealed a pattern of a rapid rise, followed by a plateau. (H) Results of the postoperative pathology test (hematoxylin and eosin staining, ×100). Cysts were formed in the intestinal submucosa, filled with mucus. Mucus overflow formed a mucus paste outside several of the cysts. No atypical gland structures or cells were observed surrounding the cysts. Fibrosis, chronic inflammatory cell infiltration, and vascular dilatation and congestion were seen in the submucosa. The patient in the control group was diagnosed with colitis cystica profunda. DCE-MRI, dynamic contrast-enhanced magnetic resonance imaging; Ktrans, volume transfer constant from the plasma compartment to the extravascular extracellular space; Kep, rate constant for transfer between extravascular extracellular space and the blood compartment; Ve, volume of extravascular extracellular space per unit volume of tissue; iAUC, initial area under enhancement curve.

Table I. DCE-MRI parameters.

Table I.

DCE-MRI parameters.

Group	n	Radiologist	Ktrans (min−1)	Kep (min−1)	Ve (%)	iAUC
Rectal cancer	40	1	0.267±0.071	0.615±0.212	0.489±0.101	37.177±8.845
		2	0.257±0.070	0.605±0.202	0.474±0.081	35.800±8.093
Controls	15	1	0.118±0.032	0.427±0.163	0.361±0.166	16.052±5.828
		2	0.121±0.027	0.411±0.160	0.367±0.148	17.328±4.428
P-valuea		1	<0.001	0.005	0.003	<0.001
		2	<0.001	0.002	0.008	<0.001

{ label (or @symbol) needed for fn[@id='tfn1-mco-0-0-762'] } Data are shown as the mean ± standard deviation.

a The Mann-Whitney U test was used to compare DCE-MRI parameters among patients with rectal cancer and controls. DCE-MRI, dynamic contrast-enhanced magnetic resonance imaging; Ktrans, volume transfer constant from the plasma compartment to the extravascular extracellular space; Kep, rate constant for transfer between extravascular extracellular space and the blood compartment; Ve, volume of extravascular extracellular space per unit volume of tissue; iAUC, initial area under enhancement curve.

The ICCs between the two independent measurements were 0.991, 0.988, 0.928 and 0.984, for Ktrans, Kep, Ve and iAUC, respectively (all P<0.001) (Table II), which suggested a good consistency of the measurements, and indicated the reliability and reproducibility of the DCE-MRI method.

Table II. Consistency of the results from the two radiologists.

Table II.

Consistency of the results from the two radiologists.

Variable	Ktrans	Kep	Ve	iAUC
ICCa	0.991	0.988	0.928	0.984
P-value	<0.001	<0.001	<0.001	<0.001

a ICC, intraclass correlation coefficient. Ktrans, volume transfer constant from the plasma compartment to the extravascular extracellular space; Kep, rate constant for transfer between extravascular extracellular space and the blood compartment; Ve, volume of extravascular extracellular space per unit volume of tissue; iAUC, initial area under enhancement curve.

ROC curve analysis of DCE-MRI parameters

ROC curve analysis was performed based on the independent measurement of the DCE-MRI parameters (Ktrans, Kep, Ve and iAUC) from the two radiologists. The areas under the ROC curves were 0.987, 0.747, 0.758 and 0.975 from radiologist 1, and 0.990, 0.777, 0.735 and 0.978 from radiologist 2, with great consistency in measurement (Fig. 3). Using a 0.156 min−1 cut-off value for Ktrans, the two radiologists determined a sensitivity of 100.0% and a specificity of 93.3%. Using a cut-off value of 24.183 or 23.410 for iAUC, the sensitivity was 92.5% for both analyses, and the specificity was 93.3% or 100.0%, respectively. A detailed analysis of the results are shown in Table III.

Figure 3. Receiver operating characteristic curves of the dynamic contrast-enhanced magnetic resonance imaging parameters measured independently by two radiologists. The AUCs of Ktrans (blue) and iAUC (red) were both >0.9, with great consistency in measurement. Ktrans, volume transfer constant from the plasma compartment to the extravascular extracellular space; Kep, rate constant for transfer between extravascular extracellular space and the blood compartment; Ve, volume of extravascular extracellular space per unit volume of tissue; (i)AUC, (initial) area under enhancement curve.

Table III. Receiver-operator characteristic (ROC) analysis for the prediction of rectal cancer using DCE-MRI.

Table III.

Receiver-operator characteristic (ROC) analysis for the prediction of rectal cancer using DCE-MRI.

	Radiologist 1				Radiologist 2
Variables	Ktrans (min−1)	Kep (min−1)	Ve (%)	iAUC	Ktrans (min−1)	Kep (min−1)	Ve (%)	iAUC
Cut-off	0.156	0.549	0.354	24.183	0.156	0.52	0.372	23.41
Sensitivity	100.00%	60.00%	92.50%	92.50%	100.00%	65.00%	85.00%	92.50%
Specificity	93.30%	86.70%	60.00%	93.30%	93.30%	86.70%	66.70%	100.00%
AUC	0.987	0.747	0.758	0.975	0.99	0.777	0.735	0.978
(95% CI)	(0.959–1.014)	(0.611–0.882)	(0.579–0.938)	(0.940–1.010)	(0.968–1.012)	(0.645–0.908)	(0.552–0.918)	(0.946–1.011)

[i] CI, confidence interval; Ktrans, volume transfer constant from the plasma compartment to the extravascular extracellular space; Kep, rate constant for transfer between extravascular extracellular space and the blood compartment; Ve, volume of extravascular extracellular space per unit volume of tissue; iAUC, initial area under enhancement curve.

Correlation between DCE-MRI parameters and pathological differentiation

The pathology reports revealed that, out of the 40 cases of rectal cancer, there were 13 poorly differentiated cases of cancer, 21 moderately differentiated cancers, and 6 well differentiated cancers. According to radiologist 1, there were significant differences in Ktrans (poorly: 0.284±0.068 and moderately: 0.280±0.067 vs. well; 0.182±0.153 min−1, P=0.004), Kep (poorly: 0.628±0.223 and moderately: 0.669±0.188 vs. well; 0.397±0.147 min−1, P=0.022), and iAUC (poorly: 41.69±4.12 and moderately: 38.17±8.45 vs. well; 23.91±3.91, P<0.001) among three groups, but not for Ve (P=0.373). Out of the four parameters, Ktrans (r=0.393, P=0.012) and iAUC (r=0.594, P<0.001) were correlated with differentiation, whereas Kep and Ve were not significantly correlated with pathological differentiation (Table IV). Similar results were obtained by radiologist 2.

Table IV. Comparison and correlation analysis between pathological differentiation and DCE-MRI parameters.

Table IV.

Comparison and correlation analysis between pathological differentiation and DCE-MRI parameters.

	Radiologist 1					Radiologist 2
Differentiation	n	Ktrans (min−1)	Kep (min−1)	Ve (%)	iAUC	n	Ktrans (min−1)	Kep (min−1)	Ve (%)	iAUC
Highly differentiated	6	0.182±0.153	0.397±0.147	0.418±0.133	23.908±3.914	6	0.175±0.015	0.393±0.143	0.422±0.077	23.059±3.700
Moderately differentiated	21	0.280±0.067	0.669±0.188	0.505±0.104	38.173±8.453	21	0.265±0.063	0.655±0.174	0.483±0.090	36.269±6.924
Poorly differentiated	13	0.284±0.068	0.628±0.223	0.496±0.067	41.691±4.115	13	0.283±0.072	0.622±0.215	0.482±0.061	40.922±4.155
P-valuea		0.004	0.022	0.373	<0.001		0.004	0.016	0.308	<0.001
rs		0.393	0.264	0.194	0.594		0.443	0.282	0.196	0.671
P-valueb		0.012	0.099	0.231	<0.001		0.004	0.078	0.227	<0.001

{ label (or @symbol) needed for fn[@id='tfn5-mco-0-0-762'] } Data are shown as the mean ± SD.

a The H test was used to compare DCE-MRI parameters among different degrees of differentiation.

b The Spearman rank correlation was used to analyze correlation between pathological differentiation and the DCE-MRI parameters. n, number; r_s, Spearman rank-order coefficient; DCE-MRI, dynamic contrast-enhanced magnetic resonance imaging; Ktrans, volume transfer constant from the plasma compartment to the extravascular extracellular space; Kep, rate constant for transfer between extravascular extracellular space and the blood compartment; Ve, volume of extravascular extracellular space per unit volume of tissue; iAUC, initial area under enhancement curve.

Discussion

Currently, the diagnosis of rectal cancer relies on traditional imaging. DCE-MRI is a relatively novel MRI technology that combines morphology and changes in hemodynamics, and can quantitatively evaluate tumor differentiation in a more accurate way (19). Therefore, the aim of the present study was to explore the clinical application of DCE-MRI in the preoperative diagnosis of rectal cancer and its correlation with tumor differentiation. The results showed that Ktrans, Kep, Ve and iAUC were higher in cancer patients compared with controls. The time-signal intensity curve of the rectal cancer lesion revealed an outflow pattern. The areas under the ROC curves for Ktrans and iAUC were both >0.9, resulting in a sensitivity and specificity of 100% and 93.3% for Ktrans, and of 92.5%, and 93.3% or 100%, for iAUC, respectively. The ICCs of Ktrans, Kep, Ve, and iAUC measured by the two independent radiologists were 0.991, 0.988, 0.972 and 0.984, respectively, indicating a good consistency. In the 40 rectal cancer cases, there was a moderate correlation between Ktrans and iAUC and pathological differentiation.

Higher blood vessel density resulting from tumor proliferation, and its synergistic effect with abnormal molecular regulation within the tumor cells, lead to abnormal angiogenesis, which is represented by leakage, twisted morphology, vascular wall expansion and crosslinking (20). This abnormal morphology results in a loose connection or loss of pericytes that nourish the endothelial cells, and huge gaps between the endothelium and basement membrane, and between the basement membrane and pericytes, leading to enlargement of the gap between vascular endothelial cells and attenuation of the maturity of blood vessels, which consequently leads to high permeability and vulnerability of newborn tumor blood vessels (21).

The most commonly used DCE-MRI parameter that reflects vascular permeability is Ktrans (the volume transfer constant) (16,22). Ktrans represents the rate at which the contrast agent transfers from the blood to the interstitium, which indicates the tumor microcirculation and the surface infiltration area. In contrast, Kep, the reverse rate constant, reflects the rate at which the contrast agent transfers from the extravascular extracellular space back to the blood. Ve is the fractional extravascular leakage volume, which predominantly reflects the percentage of contrast agent in the extravascular extracellular space (16). In addition, the semi-quantitative parameter, iAUC, is associated with tumor blood influx, perfusion and interstitium, and represents the general tumor blood flow, overall perfusion and tumor interstitial space index (16).

The present study revealed that Ktrans, Kep, Ve, and iAUC were higher in patients with rectal cancer compared with controls, indicating that massive angiogenesis and abnormal vasculature enhanced the influx of contrast agent, whereas the incomplete development of vascular endothelial cells and high vascular permeability led to increased leakage of contrast agent. However, the arteriovenous connection was observed to cause a perfusion shortcut (22), which supports the results of the present study.

Both ROC curves from the two radiologists demonstrated a large AUC (>0.9) for Ktrans and iAUC, suggesting that these two parameters had high sensitivity and specificity compared with the other two parameters. In addition, the present study also revealed that Ktrans and iAUC were correlated with the pathological differentiation of rectal cancer (0.3<r<0.8), which increased when the tumor was less differentiated. This may be partially explained by the altered vascular permeability by tumor angiogenesis, as well as the fact that poorly-differentiated tumors had more cells at metaphase, which required more nutritive elements and higher blood perfusion, and therefore a higher Ktrans. Furthermore, poorly differentiated tumors exhibited a greater heterogeneity of cell morphology and histology, higher cell density and smaller interstitium.

Neither Kep nor Ve were significantly correlated with tumor differentiation. It may be surmised that Kep and Ve were associated with the composition of extravascular, extracellular space and, despite the high local vascular permeability, the composition of the extravascular, extracellular space did not markedly differ among lesions with different differentiation statuses. Further investigation is necessary to explore the clinical relevance of these two parameters.

The results of the present study suggest that DCE-MRI parameters can be used to distinguish malignant lesions from benign ones. These results are in agreement with previous studies performed with other types of solid tumors, including orbital (23), breast (24,25), head and neck (26) and pancreatic tumors (11). A previous study in colorectal cancer is consistent with the present study (27). Another previous study also demonstrated that DCE-MRI parameters may be correlated with tumor differentiation in rectal cancer (28).

The present study is not without limitations. First, solid tumors are highly heterogeneous, and intestinal lesions often have an irregular shape. Occasionally, they cannot be distinguished from the surrounding adipose tissue due to inflammation and blood vessel invasion. If the ROI is placed at the boundary of the tumor, the results may vary, and multiple measurements are required to obtain the average. Secondly, the patients were grouped according to the state of differentiation, and the sample size was small for the well-differentiated group. Finally, DEC-MRI is a relatively novel imaging modality, and it is not yet standardized. Future studies should include more patients in order to validate the DCE-MRI measurements for rectal cancer diagnosis.

In conclusion, the present study has confirmed that DCE-MRI parameters may reflect the difference in microcirculation of rectal cancer, and Ktrans and iAUC were correlated with rectal cancer differentiation, which may provide an effective preoperative diagnosis modality for rectal cancer.

References