Role of diffusion-weighted magnetic resonance imaging in detection of lymph node metastasis in rectal cancer

Background Rectal cancer is the most prevalent gastrointestinal tumor. Early diagnosis, accurate staging as well as early treatment are the keys for improving the five-year survival rate. The objective of this research is to assess the effectiveness of diffusion-weighted MRI (DWI) in identifying lymph nodes and distinguishing between benign and metastatic nodes throughout the first stage of primary rectal cancer. Results The study showed that mean ADC value was significantly higher in mucinous carcinoma (1.72 ± 0.36 × 10 –3 mm 2 /sec) than that in non-mucinous carcinoma (0.981 ± 0.276 × 10 –3 mm 2 /sec) with a cutoff value of (1.3 × 10 –3 ) mm 2 /s which was the precise value to produce high sensitivity, specificity and accuracy of 93%, 94%, and 94%, respectively. ADC analysis showed either intermediate or low signal in 49 (70%) and high signal in 21 (30%) L.Ns. Mean ADC value showed a significant reduction in malignant L.Ns (1.01 ± 0.54 × 10 –3 mm 2 /sec) compared to benign L.Ns (1.51 ± 0.51 × 10 –3 mm 2 /sec), AUC of 0.674 ( P = 0.008) and a cutoff value of 0.987 × 10 –3 mm 2 /s with sensitivity, specificity and accuracy of 44.4%, 91.2% and 67.5%, respectively. The mean L.N /tumor ratio was 1.65 ± 0.73 in benign L.Ns and 1.06 ± 0.37 in malignant L.Ns.


Background
Colorectal cancer is the most prevalent gastrointestinal tumor.Early detection, precise determination of the stage, and prompt treatment of rectal cancer are crucial for enhancing the five-year survival rate [1].
Rectal MRI is essential for assessing rectal cancer as it offers many prognostic indicators and imaging characteristics that assist in appropriate patient care [2].
The recommendations of the European Society for Medical Oncology suggest that MRI should be used as the first imaging test due to its ability to comprehensively assess both the internal and external areas of the mesorectum.MRI has exceptional contrast resolution in soft tissue and possesses outstanding multiplanar imaging capabilities, making it a very effective technique that is often used for the first assessment of rectal cancer [3].Preoperative chemoradiotherapy and complete mesorectal excision are the established therapies for locally advanced rectal cancer in order to select the optimal therapeutic approach.It is necessary to assess the state of the lymph nodes, the stage of the tumor, and the involvement of the mesorectal fascia [4].
Multiple researches have concentrated on determining the practicality of forecasting the spread of cancer to nearby lymph nodes by imaging procedures conducted before surgery [5].The assessment of lymph node metastases in rectal cancer using computed tomography (CT) mostly relies on size-related and morphological characteristics.Typically, in CT imaging, lymph nodes that have a diameter more than 10 mm are considered to be indicative of malignancy.Furthermore, the quantitative characteristics of density, area, and heterogeneity of the lesions may be used to enhance the identification of metastatic lymph nodes in rectal cancer [6].
The objective of this research is to assess the efficacy of diffusion-weighted MRI (DWI) in identifying lymph nodes and distinguishing among benign and metastatic nodes in the first stage of primary rectal cancer.

Methods
This study was done between March 2021 and August 2022 At Ain Shams University Hospitals, Radiodiagnosis department including 35 patients; 29 patients with histopathologically proven rectal carcinoma underwent surgical resection at Ain Shams University Hospitals, and 6 cases potentially free of cancer considered as control.

Inclusion criteria
Patients with histopathologically proven rectal carcinoma, with no age or gender consideration undergoing surgical resection with available pathological report.

Exclusion criteria
Patients with certain medical conditions should not undergo magnetic resonance imaging (MRI).This includes those with implanted devices that rely on electrical, magnetic, or mechanical stimulation, such as cardiac pacemakers or automated cardioverter defibrillators since the aforementioned materials may interact with these devices.Patients who have contraindications for MRI contrast agents, such as a glomerular filtration rate (GFR) < 30 mL/min/1.73m 2 or a past severe allergic response, and have been pathologically confirmed to not have rectal cancer.

Methods/Techniques of study
Every patient underwent comprehensive history taking.Patients diagnosed with rectal cancer were scheduled to have a contrast-enhanced MR imaging of the pelvis before surgical removal.All patients were examined by 1.5 Tesla MRI machine (Philips Achieva and Intera 16 channels) using external phased-array surface coil.

MRI protocol
1. Initial three plane localizer views covering the entire pelvis.2. Pre and post fat sat axial high-resolution T2 WIs 3. Sagittal, coronal T2 WIs, axial T1 WIs. 4. Post contrast axial, sagittal and coronal T1 WIs. 5. Diffusion-weighted imaging performed in the transverse plane with tri-directional diffusion gradients by using b values of 0, 400 & 800 s/mm 2 .Mean ADC of each visible node is taken by drawing an oval-shaped region of interest, including as much of the nodal parenchyma as possible.

Histopathology
After the surgical excision, the specimens were evaluated with regard to histopathology and matched with the MRI image analysis.

Statistical analysis
A combination of Microsoft Excel 2016, the statistical tool SPSS (Statistical tool for the Social Sciences), and the MedCalC program software version 19.1 will be used to collect, organize, and analyze the data.The researchers used the mean ± SD (standard deviation), the minimum and maximum values of the range to perform descriptive statistics on numerical parametric data.The median, the first and third inter-quartile range were used for numerical nonparametric data.The number and percentage were used to assess the categorical data.

Results
This prospective study was carried out on 35 patients; 29 patients with histopathologically proven rectal carcinoma underwent surgical resection at Ain Shams University Hospitals, Radiodiagnosis Department, and 6 cases potentially free of cancer considered as control.
Out of 35 patients, there were 15 males (42.9%) and 20 females (57.1%) with a male to female ratio of 0.75:1.The mean age of patients was 46.31 ± 16.13 years and ranged from 21 to 82 years.The mean age of males was 49.93 ± 17.27 years and that of females was 43.60 ± 15.09.
According to pathological results, more than half of the cases (54.3%) had adenocarcinoma grade 2, 17.1% of them had mucinous adenocarcinoma, 5.7% of them had signet ring adenocarcinoma and 2.9% had undifferentiated carcinoma.
Figure 1 shows: Diffusion and ADC value of the rectum in the studied patients.
Regarding the relation between types of pathology of rectal carcinoma and diffusion results, it is declared in Fig. 2.  Table 2 summarizes the comparison between the presence and absence of rectal masses (case versus control) regarding diffusion and ADC value.There was a significant difference between positive and negative cases of rectal masses regarding diffusion result (p < 0.001).Mean ADC value showed a significant decrease in positive rectal masses compared to negative rectal masses (p = 0.003).Table 2, Fig. 3.
For histopathologically proven benign L.Ns, the mean size was 0.40 ± 0.17 cm.where most benign L.Ns 24 (70.6%) had oval shape and 10 (29.4%) had rounded shape.All appeared having smooth margin.On the other hands, the pathologically proven malignant L.Ns in our study, the mean size was 0.97 ± 0.37 cm.More than half L.Ns 20 (55.6%) had rounded shape and 16 (44.4%)had oval shape.Irregular margin was depicted in 35 (97.2%)L.Ns and smooth margin in 1(2.8%)Table 3.
Regarding the prognostic performance for morphological criteria of the examined lymph nodes in predicting metastasis, we found that ROC curve of the rounded shape L.N had sensitivity of 55.6% & specificity of 70.6% and that of irregular L.N margin, the sensitivity and specificity were 97.2% and 100%, respectively.High DWI and low ADC map were significantly higher in malignant L.Ns compared to benign L.Ns.Mean ADC value showed significant reduction in malignant L.Ns (1.01 ± 0.54 × 10 -3 mm 2 /sec) compared to benign L.Ns (1.51 ± 0.51 × 10 -3 mm 2 /sec).In addition, there was statistically significant difference between benign and malignant L.Ns regarding Diffusion result, L.Ns size, shape, and margin.Also, there was significant difference between them regarding T2 Signal, Enhancement as well as suspicious criteria by MRI as shown in Table 4.

ROC curve analysis for the evaluation of L.N size as an indicator of L.N. metastasis:
Showed an AUC of 0.955 (P < 0.001) and a cutoff value of 0.5 cm with sensitivity, specificity and accuracy of 88.2%, 94.4% and 96%, respectively (Fig. 4).

ROC curve analysis for the evaluation of ADC Value of L.N. as an indicator for L.N. metastasis
Showed an AUC of 0.674 (P = 0.008) and a cutoff value of 0.987 × 10 -3 mm 2 /s with sensitivity, specificity and accuracy of 44.4%, 91.2% and 67.5%, respectively (Figs. 5,6).

Evaluation of the ratio of L.N. ADC value to primary tumor ADC value (L.N /T ratio)
The mean L.N/tumor ratio was 1.65 ± 0.73 in benign L.Ns and 1.06 ± 0.37 in malignant L.Ns.L.N/tumor ratio showed a significant reduction in malignant L.Ns compared to benign L.Ns (p = 0.001).

Discussion
Colorectal cancer prevalence is considerably higher in more developed countries.However, mortality rates in more developed countries is reduced, reflecting increased screening and advancements in the diagnosis, staging, and treatment of rectal cancer [7].
Nodal metastasis is considered an important prognostic marker for local recurrence as well as disease-free survival rates.Therefore, it is important to accurately predict the need for neoadjuvant chemotherapy and radiation prior to surgery [8].
The objective of this research is to assess the efficacy of diffusion-weighted MRI (DWI) in identifying lymph nodes and distinguishing among benign and metastatic nodes in the first stage of primary rectal cancer.
The incidence rates of malignant and benign lymphadenopathy in the investigation aligned with the findings of Rutegård et al. [9], who found that preoperative MRI   detected 197 mesorectal nodal formations of which 92 (47%) could be physically correlated with histology.The prevalence rates of malignant and benign lymphadenopathy in this study agreed with the findings published by Zhuang et al. [10], who documented similar rates.In total, 346 mesorectal lymph nodes were identified during imaging assessment.Out of them, 313 were verified after histological examination; while, 33 were not matched.The approach had a success rate of 90.5%.A detailed investigation of each individual node showed that 280 (89%) of them were benign; whereas, 33 (10.5%) were determined to be cancerous.They found that the median short-axis diameters of mesorectal lymph nodes, which were histopathologically matched, were 4.0 mm (range 2.0-9.0 mm) for negative nodules and 5.6 mm (range 2.5-11.8mm) for positive nodules, as shown using MRI.The MRI data on the appearance criteria of LNs yielded compatible findings to the research conducted by Miao et al. [11], which examined a total of 794 LNs for size, margins, and morphology.Based on the histological findings, it was found that 27.3% (217 out of 794) of the lymph nodes (LNs) examined were positive for metastasis (LN +).Lymph nodes (LNs) that had an oval form with a well-preserved central hilum, or those that had an oblong shape, were more likely to be negative for metastasis (LN-).On the other hand, LNs that were lobulated and irregular in shape were more likely to be positive for metastasis (LN +).There were notable disparities between benign and malignant lymph nodes in terms of diffusion results, lymph node size, shape, and margin.
The study conducted by Zhuang et al. [10] provides support for the current research.They found that the median short-axis diameters of mesorectal lymph nodes, which were matched with histological data, were 4.0 mm (range 2.0-9.0 mm) for negative nodules and 5.6 mm (range 2.5-11.8mm) for positive nodules, as determined by MRI.
The ROC curve analysis revealed that the area under the curve (AUC) for predicting lymph node (L.N.) metastasis based on L.N. size was 0.955 (P > 0.001).The cutoff value for L.N. size was determined to be 0.5cm, with corresponding sensitivity, specificity, and accuracy values of 88.2%, 94.4%, and 96%, respectively.
The findings of Abd El Samei et al. [12] corroborated the results of this study, as they indicated that the sensitivity, specificity, and accuracy were 88.89%, 94.74%, and 91.89%, respectively.
Fowler et al. [13] reported that MRI has a high sensitivity in identifying enlarged lymph nodes, but it lacks specificity in distinguishing between benign and malignant nodes with accuracies ranging from 59 to 83%.Surov et al. [14] found that ADC values were recorded for a total of 1376 lymph nodes, with 623 (45.3%) being metastatic and 754 (54.7%) being non-metastatic.The mean apparent diffusion coefficient (ADC) value of the metastatic lymph node was determined to be 1.05 × 10 -3 mm 2 /s with a range from 0.94 to 1.15.The mean apparent diffusion coefficient (ADC) value of the non-metastatic lymph node (LN) was estimated to be 1.17 (1.01, 1.33).The sensitivity and specificity values obtained were 81% and 67%, respectively.
The findings of Ge et al. [15] corroborated the results of this study as they found that a total of 67 lymph nodes were subjected to histological investigation with 24 in the non-metastatic group and 43 in the metastatic group.The average ADC values for metastatic lymph nodes were 1.17 ± 0.16 × 10 -3 mm 2 /s, which were substantially lower compared to the average ADC values of benign lymph nodes, which were 1.29 ± 0.15 × 10 -3 mm 2 /s.
Heijnen et al. [16], reported that the use of the ADC of the nodes in relation to the main rectal tumor yielded an AUC of 0.67, with a sensitivity of 75%, specificity of 61%, PPV of 20%, and NPV of 95%.
Our present study had few limitations where individual node-to-node matching between imaging and histopathology was not attempted.Mismatch can occur when multiple lymph nodes detected by MRI and then resected, since multiple enlarged lymph nodes are commonly found around area of rectal cancer.Although we done our effort to track lymph nodes, complete matching could not be possible.

Conclusions
In rectal cancer, there was a significant difference between benign and malignant L.Ns regarding diffusion result, L.Ns size, shape, and margin.This study demonstrated the effectiveness of DWI in diagnosing lymph node metastasis in colorectal cancer; true diffusion restriction was significantly noted in malignant L.Ns compared to benign L.Ns.Mean ADC value showed a significant reduction in malignant L.Ns compared to benign L.Ns.L.N /tumor ratio showed a significant reduction in malignant L.Ns compared to benign L.Ns.

Fig. 1 Fig. 2
Fig. 1 Distribution of the studied cases as regard DWI and ADC map result

Fig. 3
Fig. 3 Comparison between positive and negative rectal masses regarding ADC Value

Fig. 5 Fig. 6
Fig. 5 ROC curve for the performance of ADC Value of Lymph nodes in predicting malignant L.N

Table 1 summarizes
DWI and ADC value of lymph nodes in the studied patients.The 800 b-value DWI images showed high signal in 63 (90%) and either intermediate or low signal in 7 (10%) L.Ns.ADC analysis showed either intermediate or low signal in 49 (70%) and

Table 1
DWI and ADC Value of lymph nodes in the studied patients

Table 2
Comparison between positive and negative rectal masses (case and control) regarding Diffusion and ADC Value

Table 3
Distribution of size, shape & margin in malignant L.N

Table 4
Comparison between benign and malignant L.N regarding different parameters Fig. 4 ROC curve for the performance of L.N size in predicting malignancy shimy et al.Egypt J Radiol Nucl Med (2024) 55:146