Correlation Between MRI and Pathology in Rectal Cancer

Keywords

Rectal Cancer - TNM staging - MRI scan

Introduction

Colorectal cancer (CRC) imposes a significant burden of morbidity and mortality in Brazil, ranking as the third leading cause of cancer-related deaths. It also stands as the third most diagnosed cancer among the population in 2022.[1] The management of rectal cancer has evolved considerably over time, notably with the introduction of the total mesorectal excision technique by Heald et al.[2] (1982), which marked a milestone in rectal cancer treatment. Additionally, the adoption of neoadjuvant therapy for patients with locally advanced tumors significantly improved therapeutic outcomes.[3]

Accurate staging plays a crucial role in the treatment of rectal cancer, as the therapeutic strategies can directly influence the patient's prognosis and quality of life.[4] Current guidelines from the National Comprehensive Cancer Network (NCCN) highlight high-resolution pelvic magnetic resonance imaging (MRI) using a specific protocol for rectal cancer as the gold standard for locoregional staging.[5] [6] Furthermore, clinical staging evaluation is as critical as pathological staging, as it can be the primary associated prognostic factor.[7]

Rectal MRI enables a detailed assessment of the tumor lesion's location and size, integrity of the mesorectal fascia, extramural vascular invasion, mesorectal invasion, lymph node involvement, relationship with the sphincter complex, and involvement with adjacent structures.[8] [9] [10] With the assistance of this method and a multidisciplinary team of specialists, it is feasible to determine the need for neoadjuvant therapy in locally advanced tumors to reduce the risk of local recurrence.[7] [11]

Additionally, the administration of preoperative chemoradiotherapy has the potential to downstage the lesion, allowing for rectal preservation and facilitating a complete surgical resection (R0). Neoadjuvant therapy can even lead to a complete clinical response, defined as the absence of residual primary tumor detected by rectal examination, endoscopic examination, and imaging,[12] with total neoadjuvant therapy (TNT) demonstrating a higher rate of complete clinical response.[13]

Restaging after neoadjuvant therapy is conducted through a new MRI, establishing a tumor regression grade (mrTRG) score of 1-5, which assesses the tumor's response to chemoradiotherapy and excludes possible tumor progressions that could influence surgical plans.[14] However, some studies report low accuracy of MRI in post-neoadjuvant restaging.[15] The histopathological analysis after surgical resection of the specimen allows for pathological staging of the lesion, contributing to the definition of more precise therapeutic outcomes.

This study aims to evaluate the correlation between clinical staging by MRI and pathological staging in patients with rectal cancer in a tertiary hospital in the Federal District, Brazil.

Material and Methods

Patients and Study Design

This study was conducted according to the ethical standards required, adhering to Resolution 196/96 of the Guidelines and Norms Regulating Research Involving Human Subjects by the National Health Council.

It is a cross-sectional study of patients from the Colorectal Surgery Department at the University Hospital of Brasília. The research includes adults, age 18 or older, with a histological diagnosis of extraperitoneal rectal adenocarcinoma who were operated on at the Colorectal Surgery Department from January 2010 to December 2022. Patients whose variables of interest were not found in medical records, those who underwent urgent operations, those with recurrent tumors, and those who underwent R2 resections were excluded from this study.

Sample Collection

Data were extracted from a prospectively maintained database: demographic data including age, sex, body mass index (BMI), physical status classification according to the American Society of Anesthesiologists (ASA), and comorbidities; clinical staging; performance of neoadjuvant therapy; type of operation performed; and pathological staging. Staging was classified based on the American Joint Commission on Cancer (AJCC) colorectal cancer staging guideline.[4] For N staging, groups were divided between absence (N0) and presence (N + ) of lymph node involvement. Clinical staging was classified into cT (T0, T1, T2, T3, T4) and cN (N0, N + ) for locoregional involvement and cStage (0, I, II, III, IV). For patients who underwent neoadjuvant therapy, the last clinical staging after the completion of neoadjuvant therapy was considered. Pathological staging was classified into pT (T0, Tis, T1, T2, T3, T4) and pN (N0, N + ) for locoregional involvement, and pStage (0, I, II, III, IV).

Data Analysis

Sample distribution was assessed using the Lilliefors (Kolmogorov-Smirnov) test. Descriptive statistics for parametric data were presented as mean and standard deviation (SD). Non-parametric data were presented as median and interquartile range (IQR). The primary outcome analyzed was the correlation between clinical staging performed by MRI and pathological staging, including subgroup evaluation of primary tumor (T) and lymph node (N) stages. The agreement and reliability between clinical and pathological staging results were evaluated using the Kappa (k) and weighted Kappa (wk) tests for nominal and ordinal variables, respectively. Based on the recommendations by Landis and Koch in 1977 for evaluating the degree of reliability, Cohen's Kappa coefficient was interpreted as follows: poor <0.0; slight 0.0-0.2; fair 0.21-0.4; moderate 0.41-0.6; substantial 0.61-0.8; almost perfect 0.81-1.0.[16]

The clinical staging results were evaluated with diagnostic tests, including sensitivity, specificity, predictive values, and likelihood ratio, using pathological staging as the gold standard for T2, T3, T4, and N0, N + . All statistical analyses were performed using R Studio software (version 2023.12.0 + 369). Statistical significance was considered for results with p values < 0.05.

Results

A total of 80 patient records were analyzed, of which 76 met the criteria for this study. Demographic data are presented in [Table 1]. Among the included patients, 51.3% were female (n = 39). The median age at the time of surgery was 62 (IQR 13.5), with a minimum age of 32 and a maximum age of 85. Almost 80% of the patients were 50 or older.

Table 1 Demographic data of the patients evaluated in the study (N = 76)

Age
Median	62 (IQ 13,5)
N < 50	16 (21,1%)
N ≥ 50	60 (78,9%)
Sex
Female	39 (51,3%)
Male	37 (48,7%)
Comorbidities
HTN	37 (48,7%)
DM II	19 (25%)
Lifestyle
Smokers	26 (34,2%)
Alcohol Consumption	5 (6,6%)
Body Mass Index (BMI)
Median	25,5 (IQ 5,2)
Underweight (BMI < 18,5)	3 (3,9%)
Normal (BMI 18,5 - 24,9)	29 (38,2%)
Overweight (BMI 25 - 29,9)	34 (44,7%)
Obese Class I (BMI 30 - 34,9)	5 (6,6%)
Obese Class II and III (BMI ≥ 35)	5 (6,6%)
American Society of Anesthesiologists (ASA) Classification
I	0 (0%)
II	56 (73,7%)
III	19 (25%)
IV	1 (1,3%)
Clinical Staging
0	1 (1,3%)
1	7 (9,3%)
2	21 (28%)
3	44 (58,7%)
4	3 (4%)
Types of Treatments
Surgery without neoadjuvant therapy	9 (11,8%)
Neoadjuvant therapy + Surgery	67 (88,2%)
Types of Surgeries
Rectum anterior resection	41 (53,9%)
Abdominoperineal amputation of the rectum	31 (40,8%)
Total proctocolectomy	3 (3,9%)
Posterior pelvic exenteration	1 (1,3%)
Pathological Staging
0	7 (9,2%)
1	10 (13,2)
2	23 (30,3%)
3	31 (40,8%)

Regarding lifestyle habits and comorbidities, 26 patients were smokers (34.2%), while five patients reported alcohol consumption (6.6%). Hypertension (HTN) was evident in 48.7% of cases, and 25% of patients had type II diabetes mellitus (DM II). The median BMI was 25.5 kg/m² (IQR 5.2), ranging from 16.8 to 46.7 kg/m². According to the ASA physical status classification, 73.7% of the individuals were categorized as ASA II, 25% ASA III, and 1.3% ASA IV.

Among the 76 patients analyzed, 88.15% (n = 67) underwent neoadjuvant therapy followed by surgery, while the other nine patients were directly subjected to surgery without prior chemoradiotherapy. Rectum anterior resection with total mesorectal excision was the most frequently performed operation, accounting for 41 surgeries (53.9%).

Regarding the last clinical T staging by preoperative MRI, 1.3% (n = 1) of tumors were classified as T0, 17.1% (n = 13) as T2, 68.4% (n = 52) as T3, and 13.2% (n = 10) as T4. The pathological staging was T0 in 7.9% (n = 6); Tis in 2.6% (n = 2); T1 in 1.3% (n = 1); T2 in 18.4% (n = 14); T3 in 65.8% (n = 50); and T4 in 3.9% (n = 3) ([Table 2]).

Table 2 Comparison and agreement between clinical and pathological staging of T

cT	pT						Total
	T0	Tis	T1	T2	T3	T4
T0	0	0	0	0	1	0	1
Tis	0	0	0	0	0	0	0
T1	0	0	0	0	0	0	0
T2	3	0	0	6 (28,5%)	4	0	13
T3	3	2	1	5	41 (67,2%)	0	52
T4	0	0	0	3	4	3 (30%)	10
Total	6	2	1	14	50	3	76

The weighted Kappa test showed fair reliability between clinical T staging and pathological T staging considering all patients in the study (kp= 0.22; p < 0.005; agreement = 65.8%). Evaluating the variables individually, there was zero agreement for Tis, T0, and T1, 28.5% for T2, 67.2% for T3, and 30% for T4.

Separately analyzing T staging of patients not subjected to neoadjuvant therapy revealed a moderate reliability weighted Kappa (kp = 0.45; p < 0.05; agreement = 66.7%). The reliability was slight for patients who underwent neoadjuvant therapy (kp = 0.19; p < 0.05; agreement = 65.7%). Evaluating each variable separately, there was zero agreement for patients not subjected to neoadjuvant therapy for Tis, T0, T1, and T4; 40% for T2; and 66.7% for T3. For patients undergoing neoadjuvant therapy, the agreement was 25%, 67.3%, and 30% for T2, T3, and T4, respectively, and zero for Tis, T0, and T1.

The sensitivity of MRI in preoperative T staging was 43% for T2, 82% for T3, and 100% for T4. For specificity, the parameters found were 89% for T2, 58% for T3, and 90% for T4. The positive predictive value for T2 was 46%, for T3 was 79%, and for T4 was 30%. The negative predictive value for T2 was 87.3%, for T3 was 62.5%, and for T4 was 100%. The positive likelihood ratio for T2 was 3.8, for T3 was 1.94, and for T4 was 10.4. The negative likelihood ratio for T2 was 0.64, and for T3 was 0.31.

Regarding the last clinical N staging by preoperative MRI, 38.2% (n = 29) of tumors were classified as N0 and 61.8% (n = 47) as N + . The pathological staging was N0 in 55.3% (n = 42) and N+ in 44.7% (n = 34) ([Table 3]).

Table 3 Comparison and agreement between clinical and pathological staging of N

cN	pN		Total
	N0	N+
N0	23 (47,9%)	6	29
N+	19	28 (52,8%)	47
Total	42	34	76

The Kappa test showed fair reliability between clinical N staging and pathological N staging considering all patients in the study (k = 0.36; p < 0.005; agreement = 67.1%). Separately analyzing N staging of patients not subjected to neoadjuvant therapy revealed substantial reliability kappa (k = 0.73; p < 0.05; agreement = 88.9%). The reliability was fair for patients who underwent neoadjuvant therapy (k = 0.29; p < 0.05; agreement = 64.2%).

MRI's sensitivity in preoperative lymph node staging was 82.35%, and its specificity was 54.76%. The positive predictive value was 59.57%, and the negative predictive value was 79.31%. The positive likelihood ratio was 1.8, while the negative likelihood ratio was 0.3.

Discussion

The patient profile in this study suggests a predominantly elderly sample with a balanced distribution between genders. Additionally, most patients underwent neoadjuvant therapy followed by surgery.

There was a varied distribution of tumors regarding clinical T staging, with a predominance of T3 (68.4%) followed by T2 (17.1%). When compared with pathological staging, the two had fair concordance (wk = 0.22). In a pivotal study, Brown et al.[9] (2003) demonstrated that the use of high-resolution MRI for clinical staging showed a 94% weighted agreement with pathological staging for T staging.[9] Brown et al.[17] (1999). demonstrated 100% concordance in T staging of 28 primary rectal cancer patients using high-resolution MRI,[17] whereas Poon et al.[18] (2005) reported an overall accuracy of 74% using a similar technique.[18] Other similar studies show a variation of 31% to 100% in overall T staging concordance,[19] suggesting that MRI may not be entirely accurate in determining the final tumor stage.

Regarding clinical N staging, most tumors were classified as N+ (61.8%) by MRI, with an overall concordance of 67.1% and fair reliability between clinical and pathological staging (k = 0.36). In a study of 354 patients who had surgery for rectal cancer, it was found that there was a 78.2% agreement for T, which has substantial reliability (wk = 0.62), and a 69.5% agreement for N with fair reliability (wk = 0.4).[19] In a study conducted at a Portuguese hospital, they found concordances of 43.9% for T and 61% for N, both with slight reliabilities described (wk = 0.14 and 0.16, respectively).[20]

Evaluating patients who underwent neoadjuvant therapy and those who did not separately, moderate reliability was observed in clinical T staging (wk = 0.45) with 66.7% concordance in patients not subjected to neoadjuvant therapy. In contrast, slight reliability was found in patients who underwent neoadjuvant therapy (wk = 0.19) with 65.7% concordance. For N staging, substantial reliability (wk = 0.73) with 88.9% concordance was noted in patients not subjected to neoadjuvant therapy. In comparison, fair reliability (wk = 0.29) with 64.2% concordance was observed in those who underwent neoadjuvant therapy. This suggests that neoadjuvant therapy may affect MRI accuracy in determining tumor stage and nodal involvement. Up to 15% of perirectal lymph nodes are too small to be identified by MRI, making lymph node metastasis detection considerably difficult.[21]

A meta-analysis of 12 articles on restaging after neoadjuvant therapy found MRI concordance for T at 52% and N at 72%, with overestimation and underestimation occurring in 10-15% of cases. The meta-analysis also showed that MRI concordance varies for different pathological T stages and is higher for T3 tumors,[15] corroborating the present study's finding of a 67.3% T3 concordance. A study by the Cleveland Clinic evaluating the reliability of MRI restaging after neoadjuvant therapy found a T concordance of 70.63% with a kappa of 0.31, suggesting fair reliability.[3]

Differentiating tumors by MRI can be challenging due to secondary changes induced by preoperative chemoradiotherapy, including edema, mucus presence, and fibrosis, which can be confused with residual tumor.[22] Additionally, the desmoplastic response of some tumors can lead to misdiagnosing T2 tumors as T3.[19]

The reliability of MRI clinical staging for both tumor size and nodal involvement was lower in patients who underwent neoadjuvant therapy. This suggests prior therapy may impact MRI accuracy in determining tumor stage and nodal involvement. This finding aligns with literature indicating that MRI concordance in rectal cancer staging post-radiotherapy appears to decrease.[23] [24]

In terms of MRI sensitivity, it was higher for T3 (82%) and T4 (100%) but lower for T2 (43%), indicating that MRI is more effective at detecting more advanced tumors. Regarding MRI specificity for each T stage, values were 89% for T2, 58% for T3, and 90% for T4, indicating a high rate of correct identification of negative cases for T2 and T4 but a relatively lower rate for T3.

Evaluating MRI sensitivity and specificity in nodal staging, sensitivity was 82%, and specificity was 54%, suggesting that MRI is more effective at detecting nodal involvement than ruling out its absence. However, Gefen et al.[3] (2023) found a sensitivity for N of 27.03% and a specificity of 82.95%.[3]

Notably, MRI's diagnostic sensitivity and specificity also largely depend on the radiologist's experience, meaning results can vary significantly between institutions and among examiners within the same institution.[25]

For T staging, the positive likelihood ratio (LR + ) indicates how much a positive MRI result increases the probability of tumor presence. In this study, the LR+ for T4 was 10.4, indicating a substantial increase in the probability of true positives in more advanced cases.

This study has limitations, primarily due to its retrospective nature. MRI exams were conducted without a standardized protocol dedicated to rectal cancer staging or the use of high-resolution studies. Additionally, reports were prepared by radiologists from different imaging centers with varying levels of experience in rectal cancer staging, which may impact result accuracy. Furthermore, the interval between the MRI and surgery was not evaluated, which could have influenced the degree of concordance.

In summary, these data highlight the utility of MRI in preoperative staging of rectal cancer but also underscore its limitations, particularly for less advanced tumors (T2) and in ruling out nodal involvement. Neoadjuvant therapy may affect MRI accuracy, emphasizing the importance of a multidisciplinary approach in evaluation, considering not only MRI results but also other factors such as proctological examination with digital rectal exam and endoscopic assessments for a precise disease staging and the development of the most appropriate treatment plan for each patient.

Conclusion

Although MRI has demonstrated good sensitivity for more advanced tumor stages, such as T3 and T4, and satisfactory specificity for T2 and T4, it showed more significant limitations in analyzing less advanced tumors (T2) and the precise evaluation of lymph nodes. MRI is more effective at detecting nodal involvement than ruling out its absence. Additionally, neoadjuvant therapy appears to affect MRI accuracy, highlighting the need for dedicated subspecialized radiologists and a multidisciplinary approach in the staging evaluation of rectal cancer.

• References

• 1 Ferlay J, Ervik M, Lam F. et al. (2024). Global Cancer Observatory: Cancer Today. Lyon, France: International Agency for Research on Cancer. Available from: https://gco.iarc.who.int/today , accessed [04 June 2024] .
• 2 Heald RJ, Husband EM, Ryall RDH. The mesorectum in rectal cancer surgery–the clue to pelvic recurrence?. Br J Surg 1982; 69 (10) 613-616
• 3 Gefen R, Garoufalia Z, Horesh N. et al. How reliable is restaging MRI after neoadjuvant therapy in rectal cancer?. Colorectal Dis 2023; 25 (08) 1631-1637
• 4 Steele SR, Hull TL, Hyman N. et al. The ASCRS Textbook of Colon and Rectal Surgery. Springer International Publishing, Imprint Springer; 2022
• 5 National Comprehensive Cancer Network. NCCN - evidence-based cancer guidelines, oncology drug compendium, oncology continuing medical education. Nccn.org. Published 2019. https://www.nccn.org/
• 6 Glynne-Jones R, Wyrwicz L, Tiret E. et al; ESMO Guidelines Committee. Rectal cancer: ESMO Clinical Practice Guidelines for diagnosis, treatment and follow-up. Ann Oncol 2017; 28 (Suppl. 04) iv22-iv40
• 7 Papamichael D, Audisio RA, Glimelius B. et al. Treatment of colorectal cancer in older patients: International Society of Geriatric Oncology (SIOG) consensus recommendations 2013. Ann Oncol 2015; 26 (03) 463-476
• 8 Smith NJ, Barbachano Y, Norman AR, Swift RI, Abulafi AM, Brown G. Prognostic significance of magnetic resonance imaging-detected extramural vascular invasion in rectal cancer. Br J Surg 2008; 95 (02) 229-236
• 9 Brown G, Radcliffe AG, Newcombe RG, Dallimore NS, Bourne MW, Williams GT. Preoperative assessment of prognostic factors in rectal cancer using high-resolution magnetic resonance imaging. Br J Surg 2003; 90 (03) 355-364
• 10 Adams AM, Vreeland TJ, Teshome M. et al. American College of Surgeons Commission on Cancer Standard 5.7 for Total Mesorectal Excision for Mid-to-Low Rectal Cancer. J Am Coll Surg 2022; 234 (06) 1249-1253
• 11 Sauer R, Liersch T, Merkel S. et al. Preoperative versus postoperative chemoradiotherapy for locally advanced rectal cancer: results of the German CAO/ARO/AIO-94 randomized phase III trial after a median follow-up of 11 years. J Clin Oncol 2012; 30 (16) 1926-1933
• 12 Habr-Gama A, Perez RO, Nadalin W. et al. Operative versus nonoperative treatment for stage 0 distal rectal cancer following chemoradiation therapy: long-term results. Ann Surg 2004; 240 (04) 711-717 , discussion 717–718
• 13 Petrelli F, Trevisan F, Cabiddu M. et al. Total Neoadjuvant Therapy in Rectal Cancer: A Systematic Review and Meta-analysis of Treatment Outcomes. Ann Surg 2020; 271 (03) 440-448
• 14 Patel UB, Taylor F, Blomqvist L. et al. Magnetic resonance imaging-detected tumor response for locally advanced rectal cancer predicts survival outcomes: MERCURY experience. J Clin Oncol 2011; 29 (28) 3753-3760
• 15 Memon S, Lynch AC, Bressel M, Wise AG, Heriot AG. Systematic review and meta-analysis of the accuracy of MRI and endorectal ultrasound in the restaging and response assessment of rectal cancer following neoadjuvant therapy. Colorectal Dis 2015; 17 (09) 748-761
• 16 Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics 1977; 33 (01) 159-174
• 17 Brown G, Richards CJ, Newcombe RG. et al. Rectal carcinoma: thin-section MR imaging for staging in 28 patients. Radiology 1999; 211 (01) 215-222
• 18 Poon FW, McDonald A, Anderson JH. et al. Accuracy of thin section magnetic resonance using phased-array pelvic coil in predicting the T-staging of rectal cancer. Eur J Radiol 2005; 53 (02) 256-262
• 19 Xu L, Zhang Z, Qin Q, Zhang C, Sun X. Assessment of T and N staging with MRI₃T in lower and middle rectal cancer and impact on clinical strategy. J Int Med Res 2020; 48 (06) 300060520928685
• 20 Abreu SFM, Martins SFF. Preoperative staging of rectal cancer with MRI: correlation with pathologic staging. J Coloproctol (Rio J) 2015; 35 (02) 77-82 . doi:https://doi.org/10.1016/j.jcol.2015.02.005
• 21 Park JS, Jang YJ, Choi GS. et al. Accuracy of preoperative MRI in predicting pathology stage in rectal cancers: node-for-node matched histopathology validation of MRI features. Dis Colon Rectum 2014; 57 (01) 32-38
• 22 Kye BH, Kim HJ, Kim G, Kim JG, Cho HM. Multimodal Assessments Are Needed for Restaging after Neoadjunvant Chemoradiation Therapy in Rectal Cancer Patients. Cancer Res Treat 2016; 48 (02) 561-566
• 23 Horvat N, Carlos Tavares Rocha C, Clemente Oliveira B, Petkovska I, Gollub MJ. MRI of Rectal Cancer: Tumor Staging, Imaging Techniques, and Management. Radiographics 2019; 39 (02) 367-387
• 24 Nahas SC, Rizkallah Nahas CS, Sparapan Marques CF. et al. Pathologic Complete Response in Rectal Cancer: Can We Detect It? Lessons Learned From a Proposed Randomized Trial of Watch-and-Wait Treatment of Rectal Cancer. Dis Colon Rectum 2016; 59 (04) 255-263
• 25 Klessen C, Rogalla P, Taupitz M. Local staging of rectal cancer: the current role of MRI. Eur Radiol 2007; 17 (02) 379-389