The Department of Radilogy of Peking University Cancer Hospital collaborated with the AI technology team to carry out top-level design and establish an AI based medical imaging management optimization path system during 2021-2022;This system can meet the needs of intelligent diagnosis and treatment,scientific research man-agement,teaching management,operation management,and assessment warning.It preliminarily demonstrates the value and role of this system in quickly providing homogeneous and high-quality diagnosis and treatment services,improving management service efficiency and quality,and improving clinical research and teaching capabilities.It is of great significance for areas with scarce high-quality medical resources and underdeveloped grassroots areas with weak medical diagnosis and treatment capabilities.The establishment of this AI based medical imaging management optimization path system provides a high-quality tool for the expansion and sinking of high-quality medical resources and regional balanced layout work,and has good prospects for promotion and application.

Magnetic resonance imaging (MRI) has been the first choice for rectal cancer due to its superb soft tissue resolution. MRI can be used to identify important factors related to treatment and prognosis, including the tumor location, TNM staging, the status of mesorectal fascia, presence of extramural vascular invasion, and involvement of the peritoneal reflection. Clinicians can make optimal surgical planning and determine the need for neoadjuvant therapy according to these factors. Patients who received neoadjuvant therapy need to undergo MRI examination after the completion of therapy with the aim of determining subsequent treatment strategies. The authors review relevant literatures, summarize the limitations and misconceptions of MRI examination in the initial staging of rectal cancer and the re-staging after neoadjuvant chemoradiotherapy, with the aim of enabling imaging physicians and clinical physicians to fully understand the advantages and limitations of MRI examination for rectal cancer, and applying MRI examination to assist in formulating clinical diagnosis and treatment strategies for rectal cancer.

Objective:To investigate the value of gray value (GV) measurement of subtraction images in contrast-enhanced spectral mammography (CESM) in the differential diagnosis of breast benign and malignant calcification.Methods:This was a retrospective study. A total of 95 patients received CESM only with mammographic calcifications without any associated mass or distortions from March 2017 to July 2021 in Peking University Cancer Hospital were enrolled. The patients were all female with an average age of 34-76(48±7) years. The craniocaudal (CC) projection of bilateral breasts was obtained prior to the mediolateral-oblique (MLO) projection. Two radiologists were asked to independently review the images to diagnose the calcification as either benign or malignant based on the presence of enhancement on subtracted imaging. GV of the calcification and background parenchyma including breast parenchyma tissue, the pectoralis major muscle and subcutaneous fatty tissue were measured by another two radiologists. The difference of lesion to background parenchyma GV (D GV) and lesion to background parenchyma gray value ratio (DR GV) were calculated. The consistency of observers was compared using the Kappa statistic. The stability of GV was evaluated with coefficient of variation. Differences of GV, D GV and DR GV between benign and malignant calcification were compared using t test or Mann-Whitney U test. Receiver operating characteristic (ROC) curves were used to analyze the efficacy of GV in differentiating benign from malignant calcification. Comparisons between diagnostic efficacies were performed using χ 2 tests. Results:Totally 97 calcification (35 malignant and 62 benign) from 95 patients were enrolled. The consistency and stability of GV values on MLO and CC projections measured by two physicians were high. The GV, D GV and DR GV of malignant calcification were significantly higher than those of the benign lesions ( P<0.05). The area under the ROC curve for the differential diagnosis of benign and malignant calcification by GV in the MLO and CC positions was 0.799 and 0.843, respectively. Threshold of calcified area GV=2015.5 in CC position, its diagnostic accuracy was 76.8%, which was similar to the subjective diagnosis of radiologists (82.1%, 78/95, P>0.05). Conclusion:Quantitative GV on subtracted imaging of CESM can differentiate benign from malignant breast calcification, especially on craniocaudal projection.

Objective:To explore the relationship between the abnormal adipose tissue around the primary tumor of colon cancer and the prognosis.Methods:From January 2015 to December 2017, 448 patients with colon cancer in Peking University Cancer Hospital were retrospectively and consecutively collected. The scores were assigned to the severity, horizontal and vertical ranges of peritumoral adipose tissue (PAT) abnormalities, and the cumulative scores were calculated to establish the PAT grades from 1 to 3. We defined a score of 0 or 1 as PAT grade 1, a score of 2 as PAT grade 2, a score of 3 as PAT grade 3. The patients were followed up, and the overall survival (OS) and metastasis-free survival (MFS) were recorded. The Kaplan-Meier curve, log-rank test and Cox regression analysis were used to evaluate its impact on prognosis, and the hazard ratio (HR) and 95% confidence interval (CI) were calculated.Results:Among the 448 patients, patients with PAT grade 1, 2, and 3 accounted for 70.1% (314/448), 18.1% (81/448), and 11.8% (53/448), respectively. The Kaplan-Meier survival curve showed that patients with PAT grade 1 had the best OS, patients with grade 3 had the worst OS, and patients with grade 2 were in between, and the difference was statistically significant (χ 2=27.38, P<0.001). There were statistically significant differences between the grades in pairs ( P<0.05). There was no significant difference in MFS between different PAT grades (χ 2=2.85, P=0.240). The results of Cox regression analysis showed that PAT grade was an independent factor affecting the OS. Compared with PAT grade 1, the risk of death in patients with PAT grade 2 and 3 was significantly increased (HR 2.563, 95%CI 1.181-5.561; HR 2.269, 95%CI 1.005-5.121; P=0.034). PAT grade was not an independent factor of MFS ( P=0.253). Conclusion:The PAT grade established based on the degree and scope of abnormal PAT in colon cancer is an independent factor for poor prognosis of colon cancer.

Objective:To investigate the efficacy of CT imaging features in evaluating occult peritoneal metastasis (OPM) of diffuse infiltrating gastric cancer (Borrmann Type Ⅳ).Methods:Totally 101 patients with locally advanced Borrmann type Ⅳ gastric cancer were retrospectively collected who were admitted to Peking University Cancer Hospital from March 2014 to March 2021. The patients were divided into OPM group (53 cases) and the non-OPM group (48 cases) according to the results of preoperative CT and laparoscopic exploration/peritoneal cytology examination. The pathological examination results were recorded, including the degree of histological differentiation and Lauren classification. The evaluation indicators included the tumor center position, the number of tumor-occupied portions, involved orientation, mucosal broadband sign, stratified enhancement, serosa invasion, increased density of peripheral fat tissue, and enlarged lymph nodes. The maximum thickness of the primary tumor, average CT value of the primary tumor (arterial phase, venous phase, and delayed phase), difference between venous phase and arterial phase, difference between delayed phase and venous phase, and pattern of the enhanced curve were recorded. The Mann-Whitney U or Chi-square test was used to compare the differences of pathological and CT features between two groups. The multivariate logistic regression was used to screen independent predictors and establish a nomogram. The receiver operating characteristic curve was used to evaluate the performance of the nomogram in predicting OPM, and the Hosmer-Lemeshow test was used to test the model′s goodness of fit. Results:There was statistical significance in the seven indicators between the OPM and non-OPM groups, including tumor-occupied portions of stomach, mucosal broadband sign, stratified enhancement, serosa invasion, increased density of peripheral fat tissue, the enhanced curve pattern and the degree of histological differentiation ( P<0.05). Among them, the degree of histological differentiation (OR=0.19, P=0.033), stratified enhancement (OR=7.02, P=0.005) and serosa invasion (OR=14.27, P<0.001) were independent predictors of OPM. The nomogram was established based on the three significant features. The area under the curve for predicting OPM was 0.826 (95%CI 0.745-0.908), the sensitivity was 0.566 and the specificity was 0.938. The Hosmer-Lemeshow test showed a good agreement between the OPM risk predicted by the nomogram and the actual risk ( P=0.525). Conclusions:The CT features of Borrmann type Ⅳgastric cancer complicated with OPM have specific characteristics. The diagnosis model based on the degree of histological differentiation, stratified enhancement, and serosa invasion had high efficacy in evaluating OPM.

Objective:To investigate the value of dynamic contrast-enhanced MRI enhancement amplitude for qualitative diagnosis of suspicious residual enhancing lesions after neoadjuvant therapy (NAT) in breast cancer.Methods:In total, 168 suspicious residual enhancing lesions of 168 patients who received NAT at Peking University Cancer Hospital from January 2015 to June 2016 were retrospectively analyzed and divided into non-residual cancer group ( n=59) and residual cancer group ( n=109) according to pathological findings. Then 168 suspicious residual enhancing lesions were stratified according to molecular subtype and baseline enhancing morphology. According to the breast imaging reporting and data system, the morphology of enhancing lesions, the margin of mass-like enhancing lesions, and the distribution of non-mass-like enhancing lesions on MRI before NAT were recorded. The second phase (1 min 45 s-2 min after contrast injection) was used as the early phase, and the fifth phase (5-6 min after contrast injection) was used as the late phase to measure the signal intensity and time-signal intensity curve (TIC) of suspicious residual enhancing lesions, and the signal enhancement ratio (SER) was calculated. Independent sample t-test, Mann-Whitney U test and χ 2 test were used to compare the difference of SER and clinical features between the non-residual and residual cancer groups. The receiver operator characteristic curve was used to analyze the diagnostic efficacy of SER to determine residual cancer. Results:There are statistically significant differences in invasive ductal carcinoma grade, hormone receptor status, the morphology of enhancing lesion on baseline MRI and TIC type between non-residual and residual cancer groups ( P<0.05). The SER values of the non-residual cancer group in the early [31% (23%, 61%)] and late (72%±43%) enhanced phases were significantly lower than those of the residual cancer group [49% (28%, 71%), 88%±38%, Z=-2.26, t=-2.43, P=0.024, 0.016, respectively]. Among suspicious residual enhancing lesions with hormone receptor negative status and single mass-like morphology, the SER values of the non-residual cancer group in the early (33%±16%) and late [64% (42%, 74%)] enhanced phases were significantly lower than those of the residual cancer group [59%±30%, 84% (77%, 106%), t=-2.86, Z=-3.17, P=0.008, 0.001, respectively]. The area under the curve values of SER in differentiating suspicious residual enhancing lesions were statistically different between early and late enhanced phases (0.606 and 0.637, respectively, Z=2.16, P=0.031). Conclusion:For breast cancer after NAT, it is difficult to determine the suspicious residual enhancing lesions on MRI subjectively, especially the hormone receptor negative lesions with single mass, SER can be used as an auxiliary diagnostic method, and it is necessary for the analysis of late enhancement.

Objective:To analyze the role of baseline mesorectal fascia (MRF) status and the correlation between MRF changes and prognosis after neoadjuvant therapy in patients with locally advanced rectal cancer.Methods:Totally 321 patients with locally advanced rectal cancer were retrospectively analyzed from January 2014 to December 2016 in Peking University Cancer Hospital. All patients underwent surgery after neoadjuvant radiotherapy and chemotherapy, and were followed up regularly after surgery. The MRF status, extramural vascular invasion (EMVI) status, tumor location, tumor stage and lymph node status were evaluated on baseline MRI. For patients with positive baseline MRF, preoperative MRF status was also evaluated. Chi-square test or independent t test were used to compare the characteristics between MRF positive and negative patients. Kaplan-Meier curve, log-rank test and multivariate Cox regression were used to analyze the correlation between imaging features and prognosis. Results:In all of the 321 subjects, 193 (60.1%) had positive baseline MRF, 54 (28.0%) of the 193 patiens had negative MRF after neoadjuvant therapy, and 139 (72.0%) of them still had positive MRF preoperatively. The postoperative pathological T and N stages were significantly higher in patients with positive baseline MRF than those with negative MRF, and the proportion of patients achieving complete pathological response was significantly lower than those with negative MRF (all P<0.05). The postoperative pathological T and N stages of patients with MRF negative conversion were significantly lower than those without MRF negative conversion. In patients with negative baseline MRF and patients with negative MRF conversion after neoadjuvant therapy, the proportion of positive MRI EMVI was significantly lower (all P<0.05). Univariate survival analysis showed that overall survival and metastasis free survival were poorer in patients with positive MRF at baseline, with a hazard ratio of 3.33 and 1.69, respectively. There was no significant correlation between negative MRF conversion after neoadjuvant therapy and overall survival, metastasis free survival and recurrence free survival. Multivariate Cox analysis showed that baseline MRF and EMVI status were independent factors for overall survival and metastasis free survival, with a risk ratio of 2.15 and 3.35 for overall survival, 1.13 and 2.74 for metastasis free survival, respectively. Conclusions:Baseline MRF status is one of the independent prognostic predictors in locally advanced rectal cancer patients with neoadjuvant therapy. However, the role of the change in MRF status after neoadjuvant therapy is uncertain for predicting prognosis.

Objective: To compare the characteristics of MRI signals in different therapeutic effect of desmoid-type fibromatosis (DF). Methods: One hundred and twenty-four DF patients with pathologically proven postoperative recurrent lesions from Peking University Cancer Hospital from 2008 to 2015 were enrolled in the study. All patients had baseline MRI scans, followed by once MRI scan at least every six months, and the follow-up period was three years. All patients had MRI images at the end of the third year. The therapeutic effect was evaluated by response evaluation criteria in solid tumors (RECIST) criteria, and the patients were divided into three groups: progressive disease (PD), stable disease (SD) and partial response (PR). Differences in features in the PD, SD, and PR groups were compared using one-way analysis of variance, Kruskal-Wallis, or Chi-square test. Multiple comparisons were performed using Bonferroni to correct P values. The spearman correlation coefficient was used to test the correlation between signal score and tumor maximum diameter. Results: Among the 124 patients, 17 experienced PD, 37 exhibited SD and 70 exhibited PR. There were no significant differences in age, gender distribution and the location of lesion among three groups (P>0.05). The difference of treatment strategies was statistically significant (P<0.05). The difference of tumor maximum diameter of baseline and the last follow-up MRI scan was also statistically significant between three groups (P<0.05). There was no significant differences in T₂ scores and enhancement scores of baseline MRI scan among the PD, SD, and PR groups (P>0.05). At the last follow-up MRI scan, the T₂ signal scores, the changes of T₂ signal, the scores of enhancement and the changes of enhancement of the PD, SD, and PR groups were significantly different among three groups (P<0.05). At the last follow-up MRI scan, the tumor maximum diameter was positively correlated with the score of T₂ signal (r=0.434, P<0.01), and the tumor maximum diameter was positively correlated with the score of enhancement (r=0.743, P<0.01). Conclusion MRI has great value in evaluating the therapeutic effect of DF.

Objective:To investigate the computed tomography (CT) features of adenocarcinoma of esophagogastric junction (AEG) after neoadjuvant chemotherapy.Methods:The retrospective and descriptive study was conducted. The clinicopathological data of 59 patients with AEG who underwent neoadjuvant chemotherapy in Peking University Cancer Hospital from February 2010 to November 2014 were collected. There were 51 males and 8 females, aged from 46 to 82 years, with a median age of 63 years. All the 59 patients underwent enhanced CT examination before and after neoadjuvant chemotherapy. Observation indicators: (1) pathological examination and neoadjuvant chemotherapy of patients with AEG; (2) results of CT examination in patients with AEG, including ① qualitative indicators of CT and ② quantitative indicators of CT. Measurement data with skewed distribution were represented as M( P25, P75) or M (range), and comparison between groups was analyzed using the Mann-Whitney U test. Count data were described as absolute numbers, and comparison between groups was analyzed by the chi-square test. Results:(1) Pathological examination and neoadjuvant chemotherapy of patients with AEG: of the 59 patients with AEG, high-differentiated adenocarcinoma was observed in 1 patient, moderate-differentiated adenocarcinoma in 40 patients, and low-differentiated adenocarcinoma in 18 patients. Effective response to neoadjuvant chemotherapy was observed in 13 patients, including 6 patients of pathological tumor regression grading (pTRG) 0 and 7 of pTRG 1; poor response was observed in 46 patients, including 12 patients of pTRG 2 and 34 patients of pTRG 3. (2) Results of CT examination in patients with AEG. ① Qualitative indicators of CT: for the 13 patients with effective response to neoadjuvant chemotherapy, 13 had the presence of ulcers, 5 had layered enhancement, 10 had infiltration of adventitia surface, and 2 had positive extramural venous invasion (EMVI) before neoadjuvant chemotherapy; after neoadjuvant chemotherapy, 13 had shallowed or disappeared ulcers, 7 patients had changed enhancement pattern, 3 had infiltration of adventitia surface, and 1 had positive EMVI. For the 46 patients with poor response to neoadjuvant chemotherapy, 28 had the presence of ulcers, 18 had layered enhancement, 37 had infiltration of adventitia surface, and 22 had positive EMVI before neoadjuvant chemotherapy; after neoadjuvant chemotherapy, 23 had shallowed or disappeared ulcers, 7 patients had changed layered enhancement pattern, 33 had infiltration of adventitia surface and 21 had positive EMVI, respectively. There was no significant difference in the layered enhancement or infiltration of adventitia surface before neoadjuvant chemotherapy between patients with different treatment response ( χ2=0.002, 0.000, P>0.05). There were significant differences in the presence of ulcers and positive EMVI before neoadjuvant chemotherapy between patients with different treatment response ( χ2=5.591, 4.421, P<0.05). After neoadjuvant chemotherapy, there were significant differences in the changes of layered enhancement pattern, infiltration of adventitia surface and positive EMVI between patients with different treatment response ( χ2=6.359, 10.090, 4.728, P<0.05); while there was no significant difference in the shallowed or disappeared ulcers between patients with different treatment response ( χ2=1.239, P>0.05). ② Quantitative indicators of CT: for the 13 patients with good response to neoadjuvant chemotherapy, the maximum tumor height, the maximum tumor area, enhanced CT value of the lesion before neoadjuvant chemotherapy were 1.37 cm(0.94 cm, 1.88 cm), 8.9 cm 2 (4.7 cm 2, 9.9 cm 2), 53 HU(47 HU, 63 HU), respectively. After neoadjuvant chemotherapy, the above indicators were 1.17 cm(0.79 cm, 1.29 cm), 4.4 cm 2(2.5 cm 2, 6.1 cm 2), 30 HU(25 HU, 53 HU), respectively. The change rates of the maximum tumor height, the maximum tumor area, and enhanced CT value of the lesion were -23%(-42%, 9%), -51%(-60 %, -21%), -44%(-51%, 19%), respectively. For the 46 patients with poor response to neoadjuvant chemotherapy, the maximum tumor height, the maximum tumor area, enhanced CT value of the lesion were 1.57 cm(1.21 cm, 1.96 cm), 9.4 cm 2(6.6 cm 2, 13.1 cm 2), 60 HU(53 HU, 66 HU) before neoadjuvant chemotherapy, respectively. After neoadjuvant chemotherapy, the above indicators were 1.16 cm(0.94 cm, 1.37 cm), 6.2 cm 2(4.8 cm 2, 8.1 cm 2), 55 HU(47 HU, 65 HU), respectively. The change rates of the maximum tumor height, the maximum tumor area, and enhanced CT value of the lesion were -27%(-38%, -9%), -33%(-47%, -12%), -9%(-22%, 9%), respectively. There was no significant difference in the maximum tumor height, the maximum tumor area, enhanced CT value of the lesion before neoadjuvant chemotherapy between patients with different treatment response ( Z=-1.372, -1.372, -1.331, P>0.05). There was no significant difference in the maximum tumor height after neoadjuvant chemotherapy between patients with different treatment response ( Z=-0.503, P>0.05), while there were significant differences in the maximum tumor area and CT value of the lesion ( Z=-2.743, -3.049, P<0.05). There was no significant difference in the change rate of the maximum tumor height or the maximum tumor area between patients with different treatment response ( Z=0.000, -1.481, P>0.05), while there was a significant difference in the change rate of CT value of the lesion ( Z=-3.231, P<0.05). Conclusion:Effective response of AEG to neoadjuvant chemotherapy was characterized by the changes in tumor layered enhancement pattern, reduction in the maximum tumor area, reduced CT value of the lesion, negative infiltration of adventitia surface, and negative EMVI.

After neoadjuvant chemoradiotherapy(nCRT), 15%-40% of rectal cnacers has a pathological complete response (pCR), with non-malignant cells demonstrated in histological assessment of the surgical resection specimen. Since these patients have excellent oncological outcomes, there has been a rapidly growing interest in organ preservation for those who achieve a clinical complete response (cCR), that is "Watch and Wait strategy" (W&W). One of the major challenges in the W&W of rectal cancer is the careful and precise selection of patients suitable for this approach. The published series on W&W has all used different modalities to assess response post nCRT, including the MSKCC′s three-tiered evaluation plan and Mercury′s mrTRG criteria. Except significant heterogeneous results, the evidence available comes mostly from retrospective cohort studies, furthermore, there is a lack of data of long-term outcomes. How to accurately screen pCR patients preoperatively is an important and difficult issue of clinical concern.