Abstract
Objective:
We proposed to determine whether the performance of inexperienced radiologists in determining extramural vascular invasion (EMVI) in rectal cancer on MRI can be promoted by means of targeted training.
Methods:
230 rectal cancer patients who underwent pre-operative chemoradiotherapy were included. Pre-therapy and post-therapy MR images and pathology EMVI evaluation were available for cases. 230 cases were randomly divided into 150 training cases and 80 testing cases, including 40 testing case A and 40 testing case B. Four radiologists were included for MRI EMVI evaluation, who were divided into targeted training group and non-targeted training group. The two groups evaluated testing case A at baseline, 3 month and 6 month, evaluated testing case B at 6 month. The main outcome was agreement with expert-reference for pre-therapy and post-therapy evaluation, the other outcome was accuracy with pathology for post-therapy evaluation.
Results:
After 6 months of training, targeted training group showed statistically higher agreement with expert-reference than non-targeted training group for both pre-therapy and post-therapy MRI EMVI evaluation of testing case A and testing case B, all p < 0.05. Targeted training group also showed significantly higher accuracy with pathology than non-targeted training group for post-therapy evaluation of testing case A and testing case B after 6 months of training, all p < 0.05.
Conclusion:
The diagnostic performance for MRI EMVI evaluation could be promoted by targeted training for inexperienced radiologist.
Advances in knowledge:
This study provided the first evidence that after 6 month targeted training, inexperienced radiologists demonstrated improved diagnostic performance, with a 20% increase in agreement with expert-reference for both pre-therapy and post-therapy MRI EMVI evaluation and also a 20% increase in or accuracy with pathology for post-therapy evaluation, while inexperienced radiologists could not gain obvious improvement in MRI EMVI evaluation through the same period of regular clinical practice. It indicated that targeted training may be necessary for helping inexperienced radiologist to acquire adequate experience for the MRI EMVI evaluation of rectal cancer, especially for radiologist who works in a medical unit where MRI EMVI diagnosis is uncommon.
Introduction
Extramural vascular invasion (EMVI) is defined as the presence of malignant cells within blood vessels beyond the muscularis propria in the vicinity of a primary colonic or rectal tumour, which is detected on pathological analysis of surgical resection specimens.1 It is usually found in locally advanced tumours1 and has been proved associated with higher rate of death and recurrence.2–4
MRI is the recommended method for pre-operative evaluation of rectal cancer,5,6 MRI features that correlate with pathological findings can be identified and used to evaluate EMVI both before and after neoadjuvant therapy.7–10 MRI EMVI is also identified as a prognostic factor for rectal cancer patient. A study of 517 rectal patients proved that pre-therapy positive MRI EMVI was an independent risk factor for overall survival and distant and local recurrence.11 Other studies suggested that persistent positive MRI EMVI status after pre-operative therapies was associated with worse disease-free survival.9,12,13 Therefore, pre-operative evaluation of MRI EMVI status is beneficial for managing proper treatment strategies and is also suggestive for individualised post-operative follow-up.
The evaluation of MRI EMVI is highly dependent on the radiologist’s subjective judgment. It is often difficult to accurately determine whether there is infiltration of tumour cells in small vessels due to the partial volume effect; it becomes harder for evaluation especially after pre-operative therapy, as a result of the induced distortion and collapse of blood vessels. The reported accordance of MRI-detected EMVI (MRI EMVI) compared with pathology EMVI was about 80%.9,14 A study reported excellent inter-reader agreement (κ was 0.858) for baseline MRI EMVI evaluation and good inter-reader agreement (κ was 0.604) for post-therapy MRI EMVI evaluation between two radiologists with different diagnostic experience.14 However, there has been no study to test the effect of radiologist’s expertise on the accuracy and inter-reader agreement for the evaluation of MRI EMVI status.
In this study, we proposed to determine whether the performance of inexperienced radiologists in determining EMVI in rectal cancer on MRI can be promoted by means of targeted training.
Methods
Study population
This retrospective study was approved by our institutional review board, and patient informed consent was waivered. The patients in this study were included from the prospective database of Peking University Cancer Hospital for rectal cancer. Consecutive patients from January 2011 to October 2013 were included by using the inclusion criteria as: (1) primary locally advanced rectal cancer (greater than or equal to T3 or node positive disease); (2) patients underwent pre-operative therapies followed by surgical resection; (3) available pathological report for acquiring diagnosis of pathology EMVI status; (4) available baseline and post-therapy MRI for acquiring diagnosis of MRI EMVI. Patients were excluded if they had history or co-existence of other cancer, or mucinous adenocarcinoma determined by pathology after total mesorectal excision surgery. Finally, 230 patients were identified to constitute the study population (Figure 1a).
Figure 1. .
Flow chart of the study. (a) the flow chart for patient inclusion and exclusion. (b) the flow chart for targeted training and testing.
MRI technique
All patients received MR examinations within one week before the initiation of pre-operative therapies (baseline MRI) and within one week before surgery (post-therapy MRI). The interval between completion of pre-operative therapies and post-therapy MRI was eight to ten weeks. All MR examinations were performed with a 3.0T MR scanner (Discovery 750; GE Healthcare) using an 8-channel-phased array body coil in the supine position. Without any bowel preparation, patients were injected intramuscularly with 20 mg scopolamine butylbromide 30 min prior to scanning to reduce colonic motility. The rectal MR protocol included axial, axial oblique, coronal, and sagittal T2W images, transverse T1W images, diffusion weighted images and gadolinium-enhanced T1WI (MR scanning parameters in Supplementary Material 1).
MRI EMVI review procedure
Four observers (R1, R2, R3, R4) were included in this study and they were randomly assigned to targeted training group (R1, R2) and non-targeted training group (R3, R4). The observers had 3 to 4 years of experience in pelvic MRI diagnosis, and approximately 10% of their full-time work was dedicated to MR diagnosis of pelvic tumours. Before starting the study, the observers received lectures on MRI diagnosis in rectal cancer, including lectures focusing on EMVI diagnosis. 20 typical cases (10 pathology EMVI positive cases and 10 pathology EMVI negative cases) with pathology EMVI diagnosis and both baseline and post-therapy MRI were mentored by three experienced radiologists who had more than 10 years of experience in pelvic MRI diagnosis. MRI EMVI evaluation was conducted using a 5-score grading system9 (Supplementary Material 1). Score 0–2 was considered as MRI EMVI negative, score 3–4 was considered as MRI EMVI positive.
The reference used for post-therapy MRI EMVI was pathology EMVI and for baseline MRI EMVI was consensus made by the three experienced radiologists (minority subordinating to the majority), respectively.
Baseline MRI and post-therapy MRI was retrieved for 230 included cases. 230 cases were randomly divided into 150 training cases and 80 testing cases. 80 testing cases were then averagely and randomly divided into testing case A and testing case B. First, the four observers independently evaluated the baseline and post-therapy MRI EMVI status using testing case A. They knew that patients had rectal cancer, but were blinded to personal information and pathology results. Second, besides clinical imaging diagnosis work, observers R1 and R2 (targeted training group) initiated a targeted training on MRI EMVI diagnosis during the next 4 months using the 150 training cases. Observers R3 and R4 (non-targeted training group) did not take targeted training on MRI EMVI diagnosis, but only practice in regular clinical imaging diagnosis work. 3 month and 6 month (to minimise the ‘memory effect’) after the first evaluation, the four observers re-evaluated testing case A, and 6 months after the first evaluation, they independently evaluated testing case B. All cases were presented to the observers in random order with patient information hidden (Figure 1b).
The targeted training on MRI EMVI diagnosis included: (a) the observers attended dedicated courses to the analysis of MRI EMVI in rectal cancer monthly; (b) the observers reviewed 150 training cases according to pathology EMVI results during 6 month, 75 cases in the first 3 months and 75 cases in the second 3 months; (c) the observers practised in regular clinical diagnosis, and also targeted in pelvic MRI diagnosis of rectal cancer, completing at least 30 MRI report for rectal cancer monthly; (d) there was a weekly meeting between the two observers in order to discuss the difficult cases and misdiagnosed cases of MRI EMVI in rectal cancer. Three experienced radiologists were introduced when there appeared discrepancy.
All sequences of MR images (including T2W images, T1W post-contrast images and diffusion-weighted images) were provided for the readers when they conducted MRI EMVI assessment or received targeted training. But they make MRI EMVI assessment or learn to improve the performance for assessment by using primarily T2W images (sag T2WI， axial T2WI and oax T2WI) and T1W post-contrast images as auxiliary.
Tumour stages (MRI T stages) and lymph node metastasis (MRI N status) were also recorded when assessing or reviewing the MRI EMVI status.
Reference for MRI EMVI evaluation
The main outcome was the agreement with the expert consensus reading for pre-therapy and post-therapy MRI EMVI evaluation. Three experienced radiologists assessed pre-therapy and post-therapy MRI EMVI status of all cases. If there was discrepancy, they discussed to reach a consensus.
The other outcome was the accuracy with pathological results for post-therapy MRI EMVI evaluation. Pathological reports were retrieved and reviewed for acquisition of pathology EMVI status. Two experienced pathologists independently extract pathology EMVI status from report. The pathology EMVI status was recorded as presence, absence and unclear. The pathology definition EMVI was: the presence of a rounded mass of tumour tissue within an endothelium-lined space, which was either surrounded by a rim of smooth muscle or contained red blood cells. If there lacked definite EMVI evaluation in pathological report, the case was judged as unclear pathology EMVI status. If there was discrepancy between the two pathologists, or conflicting information between pathology EMVI evaluation and other description in pathological report, the two pathologists discussed to reach a consensus. Among 230 included patients, 141 were pathology EMVI positive and 89 were pathology EMVI negative.
The pre-therapy and post-therapy MRI EMVI presentations according to pathology EMVI presentation were shown in Figure 2.
Figure 2. .
The pre-therapy and post-therapy MRI EMVI presentations according to pathology EMVI presentation. (a) A 59-year-old male, the pre-therapy (the first-line four images) and post-therapy (the second-line four images) sagittal and axial T2WI and enhanced images showed MRI EMVI positive (white arrow). HE staining photograph (the third-line two images) showed the extending of tumour (black arrow), invading the vascular structure (white arrow). (b) A 62-year-old male, the pre-therapy (at the top left) and post-therapy (at the top right) axial T2WI images showed MRI EMVI negative (white arrow). HE staining photograph (the second-line three images) showed the tumour cells (black arrow) were not contact the vascular structure (white arrow). EMVI, extramural vascular invasion.
Statistical analysis
Accuracy of post-therapy MRI EMVI status was evaluated according to pathology EMVI status, and sensitivity, specificity, positive predicting value (PPV), negative predicting value (NPV) and overall accuracy were calculated. Accuracy of pre-therapy, MRI EMVI status was evaluated in terms of agreement with expert radiologists, which was in fact equivalent of overall accuracy using the experts as the standard of reference. Inter-reader’s agreement was assessed in terms of κ's coefficient and was classified as: perfect (0.81–1.00), substantial (0.61–0.80), moderate (0.41–0.60), fair (0.21–0.40) or poor (0–0.20) agreement. The McNemar test was used to test the intraobserver and interobserver difference in diagnostic performance. Patient characteristics were compared among training cases and testing cases A and testing cases B, with chi-square test for category variables and student’s t-test for continuous variables. Bonferroni correction was used for multiple comparisons. All analysis was conducted with SPSS 22.0, p < 0.05 indicated statistical significance.
Result
Patient information
Summary of patient characteristics was listed in Table 1. Comparisons of patient characteristics among training cases, testing cases A and testing cases B showed there was no statistically significant difference among the three groups, all p > 0.05.
Table 1. .
Summary of patient characteristics
| Patient Characteristics | Training cases | Testing case A | Testing case B | P |
|---|---|---|---|---|
| Number of cases | 150 | 40 | 40 | / |
| Number of pEMVI positive/negative | 83/67 | 24/16 | 24/16 | 0.79 |
| Number of pre-therapy MRI EMVI positive/negative* | 97/53 | 29/11 | 27/13 | 0.64 |
| Age (year) | 57.5 ± 10.6 | 57.9 ± 11.5 | 56.1 ± 10.2 | 0.63 |
| Number of male/female | 87/63 | 23/17 | 23/17 | 0.99 |
| ypT 0–2/ypT 3–4 | 89/61 | 24/16 | 22/18 | 0.87 |
| ypN negative/ positive | 98/52 | 27/13 | 28/12 | 0.85 |
pEMVI, pathology extramural venous invasion; ypT, pathology tumour stage after neoadjuvant therapy; ypN, pathology lymph node status after neoadjuvant therapy.
pre-therapy MRI EMVI status assessed three experienced radiologists in consensus
Agreement with expert-reference
Three experienced radiologists assessed pre-therapy and post-therapy MRI EMVI status of 230 cases in consensus. Their post-therapy evaluation results were compared according to pathology EMVI, achieving an overall accuracy of 82.2% (189/230). (Supplementary Material 1).
In evaluating testing case A at baseline and after 3 months of training, no significant differences in agreement (overall accuracy) with expert-reference were found between R1/2 versus R3/4 for neither pre-therapy nor post-therapy MRI evaluation.
After 6 months of training, observers R1 and R2 showed statistically higher agreement with expert-reference than observers R3 and R4 for both pre-therapy (agreement with expert-reference were 80%, 82.5%, 60% and 62.5% for observers R1, R2, R3 and R4, respectively, p = 0.04) and post-therapy MRI EMVI evaluation of testing case A (agreement with expert-reference were 82.5%, 80%, 62.5% and 55% for observers R1, R2, R3 and R4, respectively, p = 0.03). (Figure 3, Table 2, Supplementary Material 1).
Figure 3. .
The line chart comparing the change in diagnostic performance between targeted training and non-targeted training group. (a) agreement with expert pre-therapy; (b) agreement with expert post-therapy; (c) accuracy with pathology post-therapy. There were four lines at each chart representing observer R1 (blue, targeted training group), observer R2 (red, targeted training group), observer R3 (yellow, non-targeted training group) and observer R4 (purple, non-targeted training group).
Table 2. .
Agreement of four observers in the evaluation of pre-therapy and post-therapy MRI EMVI according to expert-reference
| Agreement to expert-reference | ||||||
|---|---|---|---|---|---|---|
| Observer R1 | Observer R2 | Observer R3 | Observer R4 | P | ||
| Pre-therapy MRI | Baseline case A | 62.5% | 60% | 60% | 60% | 0.89 |
| 3-month case A | 72.5% | 70% | 60% | 62.5% | 0.59 | |
| 6-month case A | 80% | 82.5% | 60% | 62.5% | 0.04 | |
| 6-month case B | 85% | 82.5% | 57.5% | 65% | 0.03 | |
| Post-therapy MRI | Baseline case A | 50% | 57.5% | 57.5% | 55% | 0.91 |
| 3-month case A | 70% | 67.5% | 55% | 55% | 0.38 | |
| 6-month case A | 85% | 82.5% | 57.5% | 62.5% | 0.03 | |
| 6-month case B | 85% | 87.5% | 60% | 55% | 0.02 | |
In evaluating testing case B, observers R1 and R2 presented also showed statistically higher agreement with expert-reference than observers R3 and R4 both pre-therapy (agreement with reference were 85%, 82.5%, 57.5 and 62.5% for observers R1, R2, R3 and R4, respectively, p = 0.03) and post-therapy (agreement with reference were 87.5%, 80%, 55 and 62.5% for observers R1, R2, R3 and R4, respectively, p = 0.02). (Figure 3, Table 2, Supplementary Material 1).
Interobserver agreement
For both pre-therapy and post-therapy MRI EMVI evaluations, observer R1 and observer R2 achieved rising κ coefficient through the 6-month targeted training and substantial agreement in evaluating 6-month testing case A; in contrast, observer R3 and observer R4 produce stable κ coefficient through the 6 month regular clinical practice and fair agreement in evaluating 6-month testing case A. The κ of observer R1/R2 in 6-month testing case B were significantly higher than that of observer R3/R4, p = 0.01 and<0.01 for pre-therapy and post-therapy MRI EMVI evaluation, respectively (Table 3, Figure 4).
Table 3.
Interobserver agreement between targeted training group and non-targeted training group
| Interobserver agreement(κ coefficient) | Observer R1/R2 | Observer R3/R4 | P | |
|---|---|---|---|---|
| Pre-therapy MRI | Baseline case A | 0.39 | 0.35 | 0.82 |
| 3-month case A | 0.54 | 0.43 | 0.27 | |
| 6-month case A | 0.76 | 0.39 | 0.02 | |
| 6-month case B | 0.79 | 0.37 | 0.01 | |
| Post-therapy MRI | Baseline case A | 0.33 | 0.31 | 0.88 |
| 3-month case A | 0.66 | 0.35 | 0.09 | |
| 6-month case A | 0.78 | 0.38 | 0.02 | |
| 6-month case B | 0.72 | 0.27 | <0.01 | |
Figure 4. .
The histogram chart comparing the interobserver agreement between targeted training and non-targeted training group. The histogram chart demonstrating the pre-therapy (a) and post-therapy (b) κ coefficient between observers for baseline testing case A, 3-month testing case A, 6-month testing case A and 6-month testing case B. The blue histogram represented κ coefficient between observer R1 and observer R2; the red histogram represented κ coefficient between observer R3 and observer R4.
Correlation with pathology post-therapy
The four observers showed similar sensitivity, specificity, PPV, NPV and overall accuracy in evaluating post-therapy testing case A at baseline and after 3 months of training. Observer R1 and observer R2 showed significantly higher sensitivity and overall accuracy (82.5 and 80%) than observers R3 and R4 (62.5 and 55%) in evaluating testing case A after 6 months of training; P values were 0.02 and 0.02, respectively. Observer R1 and observer R2 showed significantly higher sensitivity specificity, PPV, NPV and overall accuracy than observers R3 and R4 in evaluating testing case B (overall accuracy for R1 and R2 were 87.5 and 80%, versus 55 and 52.5% for R3 and R4), P values were 0.05, 0.02, 0.04, 0.03 and <0.01, respectively. (Figure 3, Table 4, Supplementary Material 1).
Table 4. .
Diagnostic performance of four observers in the evaluation of post-therapy MRI EMVI according to pathology EMVI
| Diagnostic performance | Observer R1 | Observer R2 | Observer R3 | Observer R4 | P | |
|---|---|---|---|---|---|---|
| Baseline testing case A | sensitivity | 66.7% | 70.8% | 66.7% | 54.2% | 0.65 |
| specificity | 50% | 31.3% | 56.3% | 50% | 0.52 | |
| PPV | 66.7% | 60.7% | 69.6% | 61.9% | 0.49 | |
| NPV | 50% | 56.3% | 52.9% | 42.1% | 0.89 | |
| accuracy | 60% | 55% | 62.5% | 52.5% | 0.64 | |
| 3-month testing case A | sensitivity | 83.3% | 75% | 66.7% | 58.3% | 0.26 |
| specificity | 56.3% | 50% | 50% | 56.3% | 0.92 | |
| PPV | 74.1% | 69.2% | 66.7% | 60.8% | 0.79 | |
| NPV | 69.2% | 57.1% | 50% | 41.1% | 0.78 | |
| accuracy | 72.5% | 65% | 60% | 52.5% | 0.31 | |
| 6-month testing case A | sensitivity | 91.7% | 87.5% | 70.8% | 58.3% | 0.02 |
| specificity | 68.8% | 68.8% | 50% | 50% | 0.51 | |
| PPV | 81.5% | 80.8% | 68% | 63.6% | 0.38 | |
| NPV | 84.6% | 78.6% | 53.3% | 44.4% | 0.06 | |
| accuracy | 82.5% | 80% | 62.5% | 55% | 0.02 | |
| 6-month testing case B | sensitivity | 87.5% | 83.3% | 62.5% | 60.9% | 0.05 |
| specificity | 87.5% | 75% | 43.8% | 43.8% | 0.02 | |
| PPV | 91.3% | 83.3% | 62.5% | 60.9% | 0.04 | |
| NPV | 82.3% | 75% | 43.8% | 50% | 0.03 | |
| accuracy | 87.5% | 80% | 55% | 52.5% | <0.01 | |
EMVI, extramural venous invasion; NPV, negative predictive value; PPV, positive predictive value.
After 6 months of training, observer R1 achieved higher sensitivity (91.7% versus 66.7%, p = 0.03) and overall accuracy (82.5% versus 60%, p = 0.03) than baseline for evaluating case A. Observer R2 achieved higher overall accuracy (80% versus 55%, p < 0.01) than baseline. However, observer R3 and observer R4 did not present higher overall accuracy after 6 months of regular clinical practice, P values were 0.86 (62.5% versus 62.5%) and 0.82 (52.5% versus 55%) for observer R3 and observer R4, respectively.
Discussion
This study found that inexperienced radiologists after 6 months of targeted training, showed statistically higher agreement with expert-reference for both pre-therapy and post-therapy MRI EMVI evaluation compared with those without targeted training. Inexperienced radiologists after 6 months of targeted training also showed significantly higher accuracy with pathology for post-therapy evaluation. This was the first time we proved that inexperienced radiologist’s expertise for MRI EMVI evaluation could be promoted after targeted training.
Most of the studies found that with the accumulation of radiologist’s expertise as a result of targeted training, the accuracy of diagnosis is improved. Several studies focusing on the accuracy of detection and staging of prostate cancer showed that the longer time the radiologist spent in MRI diagnosis for prostate cancer, the higher accuracy they achieved.15–17 These provided strong evidence that the accuracy of detection and staging of prostate cancer are highly dependent on radiologist’s expertise. Miglioretti’s study found that radiologists’ interpretations of screening mammograms improve significantly during their first few years of practice and continue to improve throughout much of their careers. Additional residency training and targeted continuing medical education may help reduce the number of work-ups of benign lesions while maintaining high cancer detection.18 However, a study focusing on the diagnostic accuracy of multiparametric MRI for the detection of locally advanced prostate cancer prior to radical prostatectomy drew different conclusion, where the diagnostic accuracy did not improve significantly over time.19 It is worth noting that this study carried out on a hospital level which involved radiologists with different expertise, as a result it may not fit for individual radiologists.
There was a large range of incidence of pathology EMVI reported in prior studies.1,9,11,14,20–22 A likely reason for the inconsistent pathology EMVI incidence was the lack of use of elastic tissue stains, which can help distinguish lymphatic from venous invasion.23 In addition, pre-operative therapy leads to fibrosis and distortion of the normal microarchitecture which guides pathologists in identifying venous invasion, leading to a known false-negative rate during histopathological analysis. Whereas a post-therapy MRI clearly shows massive tumour extension into the extramural veins which can be followed anatomically, the cross-sectional pathology sampling does not permit the pathologist to follow the vessels in the coronal or sagittal plane, and consequently pathologists have difficulty in detecting EMVI. Presently, pre-operative MRI EMVI status is regularly evaluated in our clinical practice, as there has been evidence that baseline positive MRI EMVI is a potentially risk factor for assigning patients to further pre-operative treatment or increased follow-ups.11
Despite it is believed that the MRI EMVI evaluation corresponds well with pathology EMVI, there were few studies focusing on the diagnostic accuracy of MRI EMVI evaluation and testing the consistency among observers, the sample size is yet small.9,14 These studies reported good accuracy and interobserver agreement of experienced radiologists. Other studies about MRI EMVI evaluation often assigned two independent observers for evaluation and a third one for arbitration. There has been no study testing how the accumulated expertise of observers will affect the diagnostic performance.
This study found that after 6-month targeted training, inexperienced radiologists demonstrated a 20% increase in agreement with expert-reference for both pre-therapy and post-therapy MRI EMVI evaluation, improved interobserver agreement and also a 20% increase in or accuracy with pathology for post-therapy evaluation. Actually the ultimate overall accuracy level reached by the targeted training group post-CRT is similar to that of the expert readers (82.2% for expert, see Supplementary Material 1); while inexperienced radiologists could not gain obvious improvement in MRI EMVI evaluation through the same period of regular clinical practice. This study was carried out in a cancer hospital, and radiologists working here spends approximately 10% of their full-time work in the MR diagnosis of pelvic tumours. While in general hospitals, the proportion of such diagnosis is significantly lower. Therefore, it may be very difficult to accumulate expertise to improve the diagnostic accuracy of MRI EMVI evaluation from long-term clinical practice time. We assumed that for such uncommon diagnosis, intensive targeted training may be more effective, especially for inexperienced radiologists, and the results of this study also supported our initial hypothesis.
This study has several limitations. First, this study included only four observers in a single centrer. Our unit is a cancer hospital, thus our radiologists may obtain more experience in the diagnosis of rectal cancer from clinical practice than radiologists in general hospitals. If multiple hospitals were included, this study may include radiologists with significantly different diagnostic experience in MRI EMVI evaluation at the initiation of study. The experience obtained from previous practice is very likely to affect the diagnostic performance and the affect of targeted training. Thus, in this study, we included four observers with similar experience in MRI diagnosis of rectal cancer, and randomly assigned them to targeted training group and non-targeted training group, in order to obtain two comparable groups at baseline. Actually, the results showed that targeted training group and non-targeted training group demonstrated similar diagnostic performance at baseline. Second, our study found that, with the accumulation of radiologist’s expertise after 6 months of targeted training, targeted training group achieved better diagnostic performance, but it is uncertain if the effect of the high-intensity targeted training could be sustained for a long time in clinical practice. Third, this study only included patients undergoing pre-operative therapy, since this subject is underrepresented in current literature. As a consequence, pre-therapy EMVI readings could not be correlated with histopathology.
This study proved that diagnostic performance for MRI EMVI evaluation can be promoted significantly as a result of short-term targeted training for inexperienced radiologists. Targeted training may be necessary for helping inexperienced radiologist to acquire adequate experience for the MRI EMVI evaluation of rectal cancer, especially for radiologist who works in a medical unit where MRI EMVI diagnosis is uncommon. In future, multicentre study, or study on hospital level including both rectal cancer patients with and without pre-operative therapies will be needed to test the long-term effect of targeted training.
Footnotes
The authors Shuai Wang and Xiao-Ting Li contributed equally to the work.
Funding: This study was funded by Beijing Municipal Administration of Hospitals Clinical Medicine Development of Special Funding Support (No.ZYLX201803), the National Natural Science Foundation of China (No. 81971584), Beijing Municipal Science & Technology Commission (Z171100001017102), Beijing million Talents Project (No.2017A13) and Beijing Hospitals Authority Youth Programme (Code:QML20181103).
Ethics approval: Institutional Review Board approval was obtained from the Medical Ethics Committee of Peking University Cancer Hospital.
Competing interests: None declared
Patient consent: Patient consent was waived by the Medical Ethics Committee of Peking University Cancer Hospital
Contributors: YS.S designed the study. S.W and XT.L collected patient infornation and MR images. S.W, XY.Z and XT.L designed the training method guided the targeted training. RJ.S, YH.Q, HC.Z and Z.G conducted image measurement. XT.L conducted statistical analysis. S.W and XT.L drafted the paper, all authors edited and approved the papaer. YS.S was in charge of the paper.
Contributor Information
Shuai Wang, Email: wangshuai2011@foxmail.com.
Xiao-Ting Li, Email: lixiaoting198407@163.com.
Xiao-Yan Zhang, Email: 370493077@qq.com.
Rui-Jia Sun, Email: srj777@163.com.
Yu-Hong Qu, Email: light17@126.com.
Hui-Ci Zhu, Email: 15201664634@126.com.
Zhen Guan, Email: 18801231091@163.com.
Ying-Shi Sun, Email: sys27@163.com.
REFERENCES
- 1.Talbot IC, Ritchie S, Leighton M, Hughes AO, Bussey HJ, Morson BC. Invasion of veins by carcinoma of rectum: method of detection, histological features and significance. Histopathology 1981; 5: 141–63. doi: 10.1111/j.1365-2559.1981.tb01774.x [DOI] [PubMed] [Google Scholar]
- 2.Freedman LS, Macaskill P, Smith AN. Multivariate analysis of prognostic factors for operable rectal cancer. Lancet 1984; 2: 733–6. doi: 10.1016/S0140-6736(84)92636-9 [DOI] [PubMed] [Google Scholar]
- 3.Michelassi F, Block GE, Vannucci L, Montag A, Chappell R. A 5- to 21-year follow-up and analysis of 250 patients with rectal adenocarcinoma. Ann Surg 1988; 208: 379–89. doi: 10.1097/00000658-198809000-00016 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Horn A, Dahl O, Morild I. Venous and neural invasion as predictors of recurrence in rectal adenocarcinoma. Dis Colon Rectum 1991; 34: 798–804. doi: 10.1007/BF02051074 [DOI] [PubMed] [Google Scholar]
- 5.Beets-Tan RG, Lambregts DM, Maas M, Bipat S, Barbaro B, Caseiro-Alves F, et al. Magnetic resonance imaging for the clinical management of rectal cancer patients: recommendations from the 2012 European Society of gastrointestinal and abdominal radiology (ESGAR) consensus meeting. Eur Radiol 2013; 23: 2522–31. doi: 10.1007/s00330-013-2864-4 [DOI] [PubMed] [Google Scholar]
- 6.Glynne-Jones R, Wyrwicz L, Tiret E, Brown G, Rödel C, Cervantes A, et al. Rectal cancer: ESMO clinical practice guidelines for diagnosis, treatment and follow-up. Ann Oncol 2017; 28(suppl_4): iv22–40. doi: 10.1093/annonc/mdx224 [DOI] [PubMed] [Google Scholar]
- 7.Mercury Study Group Diagnostic accuracy of preoperative magnetic resonance imaging in predicting curative resection of rectal cancer: prospective observational study. BMJ 2006; 333: 779. doi: 10.1136/bmj.38937.646400.55 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Patel UB, Taylor F, Blomqvist L, George C, Evans H, Tekkis P, et al. Magnetic resonance imaging-detected tumor response for locally advanced rectal cancer predicts survival outcomes: MERCURY experience. J Clin Oncol 2011; 29: 3753–60. doi: 10.1200/JCO.2011.34.9068 [DOI] [PubMed] [Google Scholar]
- 9.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: 229–36. doi: 10.1002/bjs.5917 [DOI] [PubMed] [Google Scholar]
- 10.Smith NJ, Shihab O, Arnaout A, Swift RI, Brown G. MRI for detection of extramural vascular invasion in rectal cancer. AJR Am J Roentgenol 2008; 191: 1517–22. doi: 10.2214/AJR.08.1298 [DOI] [PubMed] [Google Scholar]
- 11.Zhang X-Y, Wang S, Li X-T, Wang Y-P, Shi Y-J, Wang L, et al. MRI of extramural venous invasion in locally advanced rectal cancer: relationship to tumor recurrence and overall survival. Radiology 2018; 289: 677–85. doi: 10.1148/radiol.2018172889 [DOI] [PubMed] [Google Scholar]
- 12.Chand M, Evans J, Swift RI, Tekkis PP, West NP, Stamp G, et al. The prognostic significance of postchemoradiotherapy high-resolution MRI and histopathology detected extramural venous invasion in rectal cancer. Ann Surg 2015; 261: 473–9. doi: 10.1097/SLA.0000000000000848 [DOI] [PubMed] [Google Scholar]
- 13.Chand M, Swift RI, Tekkis PP, Chau I, Brown G. Extramural venous invasion is a potential imaging predictive biomarker of neoadjuvant treatment in rectal cancer. Br J Cancer 2014; 110: 19–25. doi: 10.1038/bjc.2013.603 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Jhaveri KS, Hosseini-Nik H, Thipphavong S, Assarzadegan N, Menezes RJ, Kennedy ED, et al. MRI detection of extramural venous invasion in rectal cancer: correlation with histopathology using elastin stain. AJR Am J Roentgenol 2016; 206: 747–55. doi: 10.2214/AJR.15.15568 [DOI] [PubMed] [Google Scholar]
- 15.Harris RD, Schned AR, Heaney JA. Staging of prostate cancer with endorectal MR imaging: lessons from a learning curve. Radiographics 1995; 15: 813–29. doi: 10.1148/radiographics.15.4.7569131 [DOI] [PubMed] [Google Scholar]
- 16.Akin O, Riedl CC, Ishill NM, Moskowitz CS, Zhang J, Hricak H. Interactive dedicated training curriculum improves accuracy in the interpretation of MR imaging of prostate cancer. Eur Radiol 2010; 20: 995–1002. doi: 10.1007/s00330-009-1625-x [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Gaziev G, Wadhwa K, Barrett T, Koo BC, Gallagher FA, Serrao E, et al. Defining the learning curve for multiparametric magnetic resonance imaging (MRI) of the prostate using MRI-transrectal ultrasonography (TRUS) fusion-guided transperineal prostate biopsies as a validation tool. BJU Int 2016; 117: 80–6. doi: 10.1111/bju.12892 [DOI] [PubMed] [Google Scholar]
- 18.Miglioretti DL, Gard CC, Carney PA, Onega TL, Buist DS, Sickles EA, et al. When radiologists perform best: the learning curve in screening mammogram interpretation. Radiology 2009; 253: 632–40. doi: 10.1148/radiol.2533090070 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Jansen BHE, Oudshoorn FHK, Tijans AM, Yska MJ, Lont AP, Collette ERP, et al. Local staging with multiparametric MRI in daily clinical practice: diagnostic accuracy and evaluation of a radiologic learning curve. World J Urol 2018; 36: 1409–15. doi: 10.1007/s00345-018-2295-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Knudsen JB, Nilsson T, Sprechler M, Johansen A, Christensen N. Venous and nerve invasion as prognostic factors in postoperative survival of patients with resectable cancer of the rectum. Dis Colon Rectum 1983; 26: 613–7. doi: 10.1007/BF02552975 [DOI] [PubMed] [Google Scholar]
- 21.Minsky BD, Mies C, Rich TA, Recht A, Chaffey JT. Potentially curative surgery of colon cancer: the influence of blood vessel invasion. J Clin Oncol 1988; 6: 119–27. doi: 10.1200/JCO.1988.6.1.119 [DOI] [PubMed] [Google Scholar]
- 22.Krasna MJ, Flancbaum L, Cody RP, Shneibaum S, Ben Ari G. Vascular and neural invasion in colorectal carcinoma. Incidence and prognostic significance. Cancer 1988; 61: 1018–23. doi: [DOI] [PubMed] [Google Scholar]
- 23.Messenger DE, Driman DK, Kirsch R. Developments in the assessment of venous invasion in colorectal cancer: implications for future practice and patient outcome. Hum Pathol 2012; 43: 965–73. doi: 10.1016/j.humpath.2011.11.015 [DOI] [PubMed] [Google Scholar]