Computed Tomography as a Single Modality for Preoperative Tumor Staging for Locally Advanced Rectal Cancer. A Descriptive Tomographic Scoring System

Introduction

The current standard of preoperative staging of rectal cancer includes a rectal Magnetic Resonance Imaging (MRI) or Endorectal Ultrasound (ERUS) for evaluation of the stage of local disease, and a Computerized Tomography (CT) for assessing the stage of systemic disease [1,2]. Accurate preoperative staging determines the appropriate treatment sequence, where low stage tumors can undergo upfront surgery, and higher stage, locally advanced tumors usually require preoperative Chemoradiotherapy (CRT) [3,4]. The diagnostic accuracy of CT, MRI and ERUS for preoperative staging of rectal tumors was assessed in several studies in which the modalities had relatively comparable results [5-13]. While high-resolution MRI is generally considered superior for rectal cancer staging – particularly in assessing tumor depth, mesorectal fascia involvement and Circumferential Resection Margin (CRM) status – there are specific areas where CT shows comparable performance, notably, in early-stage tumors and nodal status [513]. MRI is superior to CT for identifying tumor penetration into the rectal wall primarily because of its superior soft tissue contrast and spatial resolution, which allow for precise differentiation of the rectal wall layers and surrounding structures. In particular, the high-resolution T2-weighted MRI sequences, provides excellent contrast between the different rectal wall layers. CT, in contrast, cannot reliably distinguish rectal wall layers due to limited softtissue contrast resolution.  CT cannot show these distinctions clearly; it often detects gross wall thickening or irregular outer margins, but not precise wall layer involvement.The role of CT for Locally Advanced Rectal Cancer (LARC) staging has been previously discussed in terms of mesorectum and mesorectal fascia invasion [7-10]. Deep invasion of mesorectum can be easily identified by CT, MRI and ERUS, while for superficial invasion it may be challenging and MRI is usually required. In such circumstances, decision making based on CT findings is not well established yet. However, due to a better availability of CT scanners and in order to rule out metastatic disease, in most cases the staging CT is performed prior to MRI and ERUS. We hypothesized that in cases in which the reported CT results described a locally advanced rectal tumor, the added value of the MRI and ERUS that might influence the treatment plan would be minimal, thus the decision to proceed with preoperative CRT could be made based on the results of the CT alone. In these cases, the MRI exam could be deferred in order to not delay treatment. Additionally, given the limited CT soft tissue contrast resolution, the radiologists usually use descriptive parameters rather than actual TNM terminology when reporting the local extent of the rectal tumor. Using standard terms in the CT report may provide better tools to determine the appropriate treatment. In an attempt to achieve this goal, we propose a simple scoring system that translates standard descriptive CT findings into numeric parameters. These descriptive findings are often discussed in the standard CT reports.

Methods

After obtaining an IRB approval following the Helsenki declaration, we performed a retrospective analysis that included all the patients with mid and low rectal cancer, defined as ≤ 10 cm from anal verge, who were treated in the department of surgery at Carmel Medical Center in Haifa, Israel from 2017 to 2020. Patients and tumor characteristics were collected. With regard to oncologic work-up, all the patients underwent colonoscopy and were routinely evaluated by CT and MRI or ERUS. Based on the radiologic findings a decision was made on preoperative CRT or upfront surgery. CT and MRI scans were performed at multiple institutions and image interpretations were reported by different senior radiologists and were reevaluated by an experienced body radiologist at our institute. ERUS was performed by a single experienced colorectal surgeon. Findings with regard to local tumor extent were retrieved from the MRI and ERUS reports. After reviewing all the CT reports, we found 4 descriptive findings that were most commonly reported by radiologists in the CT reports regarding the extent of rectal tumors: rectal wall thickening or space occupying lesion, peri-rectal fat haziness, enlarged/ prominent peri-rectal lymph nodes and penetration into adjacent organs or structures. The absence or presence of each of these findings in the CT report received a score of 0 or 1 respectively, comprising a total potential CT score between 0 (none of the findings was reported), and 4 (all of the findings were reported). Since each parameter may reflect a degree of locally advanced disease, we classified patients into two subgroups; low and high probability for advanced disease. Group A – patients with low CT scores of 0 or 1 (low probability), and Group B – patients with high CT scores of 2-4 (high probability). In all the patients, the decision on preoperative CRT was made per the standard of care according to the disease stage on MRI/ ERUS regardless of the CT scores. Patients in whom the stage on MRI or ERUS was at least T3 and/or N+ were referred for preoperative CRT.

Statistical Analysis

Continuous variables are presented by mean, and Standard Deviation (SD) or median & (interquartile range) IQR. The categorical variables are presented in percentages. Differences between the groups of high and low CT score were evaluated using Chi Square with Yates correction or Fisher exact test for the categorical variables and Independent t-test for the continuous variables. P<0.05 was considered statistically significant.

Results

Between 2017 and 2020, 80 patients were assessed and treated for mid and low rectal cancer, of whom complete data were available in 74 patients (F=49). The patients were classified into two groups according to low and high CT scores. There were 42 patients in group A (low score) and 32 patients in group B (high score). Patients’ and tumor characteristics are listed in Table 1.  Median tumor distance from anal verge was 7cm (1-10) in group A and 6cm (1-10) in group B (p-0.41). Sixty-six patients underwent low anterior resection, 5 patients - abdominoperineal resection and three patients did not undergo surgery. Postoperative tumor stage was comparable between groups (p- 0.9) (Table 1). Advanced preoperative local tumor stage did not correlate with postoperative pathology findings. Most patients were stage I or stage II tumor. In the entire cohort 13.5% experienced complete pathological response, 7 patients in the low CT score group and 3 patients in the high CT score group. In group A, out of 42 patients who had low CT scores, the tumor stage on MRI and/or ERUS was actually at least T3/N+ in 40 patients, and they were referred for CRT per the standard of care. Only in 2 patients in this group the tumor stage on MRI/ERUS correlated to the low CT score, and they were referred for upfront surgery. In group B, all 32 patients had a high CT score and locally advanced tumors on MRI/ERUS, and all of them were referred for CRT. We compared the CT results to MRI/ERUS, with the latter studies being the gold standard for decision making for preoperative CRT.

CT- Computerized Tomography, ASA – American Society of Anesthesiology, AV- Anal Verge

Table 1: Patients’ and tumor characteristics.

Discussion

The routine preoperative staging process of rectal cancer includes abdominal and chest CT and pelvic MRI. Their results often allow for a decision making on the appropriate treatment sequence. In the majority of cases, the staging CT is performed prior to the MRI/ ERUS, and in most cases the degree of local spread of the rectal tumor can be assessed based on that CT scan. Herein we showed that in selected cases, when a locally advanced rectal tumor is clearly visualized on CT, it may be reasonable to proceed with neoadjuvant chemoradiation without MRI. Examples for cases when CT is sufficient: Bulky tumor with clear extramural spread where CRM is obviously threatened; CT shows invasion into adjacent structures (unequivocal T4); patient is not a candidate for surgery and CRT is either palliative or aimed at downsizing for further decisions or when there is high clinical urgency and MRI is unavailable or contraindicated. Our hypothesis was that in such cases in which the CT shows an obvious locally advanced rectal cancer, the tomographic findings will usually suggest a high CT score, and the additional information achieved by a subsequent MRI/ERUS does not influence the immediate treatment plan of whether to refer the patient to preoperative CRT or not (Table 2). As such, the MRI/ERUS could be deferred in the immediate workup time, thus expediting the preoperative treatment. There is a definitive advantage of MRI over CT in terms of rectal cancer staging. In order to improve CT performance and accessibility in rectal cancer staging, we created a new scoring system based on routine CT findings. Since the radiological report of the CT scan uses descriptive terms rather than standard staging terminology, we designed a simple scoring system to convert descriptive wording to numeric parameters, hence a “CT score”, regarding the degree of local spread of rectal cancer – as described above. The score is based on 4 common reported findings on CT scan reports, assuming that the higher the score – the more advanced is the rectal cancer. We divided the patients into 2 groups of low and high CT scores and reviewed their treatment course that was determined by appropriate local staging by MRI/ERUS.

Interestingly, the vast majority of our patients, in both score groups (72/74), were diagnosed with locally advanced rectal cancer and had preoperative CRT. Only 2 patients, with low CT score, skiped preoperative CRT and underwent upfront surgery based on MRI/ ERUS findings. These data suggest that MRI/ERUS are probably more accurate than CT for determining the degree of local spread of rectal cancer in patients with low CT scores, in whom the CT tends to underestimate the degree of local disease, as opposed to patients with high CT scores (locally advanced tumors), where the results of the MRI/ERUS concur with the results of the CT, thus creating some form for redundancy. As such, we believe that MRI/ERUS are mandatory in patients with low CT score, i.e. no remarkable local tomographic findings, prior to treatment planning. These patients may be misdiagnosed with T2 or early T3 tumors.

CT- Computerized Tomography, MRI- Magnetic Resonance Imaging, ERUS-Endorectal Ultrasound

Table 2: Staging modalities and CT scores.