| Objective:The objective of the study is to explore the shrinkage modes of the primary tumor inwomen with breast cancer after neoadjuvant chemotherapy (NAC) with part-mountsub-serial section (PMSS) and three-dimensional (3D) pathological reconstructiontechnique, and combine clinical variables associated with the shrinkage modes into aPrediction nomogram. The second objective is to investigate the predictive value of3DMRI reconstruction for the shrinkage modes and pathologic residual tumors extent afterNAC.Methods:From June2008to August2013, Eighty-six women from Breast cancer center ofShandong cancer hospital with pathologically proven solitary invasive ductal carcinoma(Ⅱ A-Ⅲ C) were recruited. They were divided into two groups. Group A (n=25) receivedhalf cycles of NAC and Group B (n=61) received whole cycles of NAC. Breast specimenwas prepared with PMSS, and residual tumors were microscopically outlined, scanned andregistered by PHOTOSHOP software. The3D model of residual tumors was reconstructedwith3D-DOCTOR software by pathology and MRI to evaluate the shrinkage mode andmeasure the residual tumor size. A logistic regression model was used to construct thenomogram for Predicting the shrinkage mode.We defined the pathological shrinkage modes as5categories:Ⅰ. Surgical pCR: no residual tumors, Ⅱ. solitary lesion withoutsurrounding lesions, Ⅲ. multinodular lesions, Ⅳ. solitary lesion with adjacent spottylesions, and Ⅴ. diffuse lesions. Further, the clinic-pathologic shrinkage modes weredivided into2categories: concentric shrinkage mode (CSM, the longest diameter of thePathological residual tumors was less than50%and≤2cm in comparison with the primarytumor before NAC) and non-concentric shrinkage mode (NCSM, the longest diameter ofthe pathological residual tumors was more than50%and/or>2cm in comparison with theprimary tumor before NAC).Results:1. Pathological shrinkage modes：Group A: modeⅠ, Ⅱ, and Ⅴ was observed in1(4.0%),1(4.0%), and23(92.0%) cases, respectively, Group B: modeⅠ, Ⅱ, Ⅲ, Ⅳ, andⅤ was observed in18(29.5%),3(4.9%),12(19.7%),21(34.4%), and7(11.5%) cases,respectively (P=0.000). There was a statistic difference in the rate of modeⅠin two groups(P=0.009). Clinic-pathologic shrinkage modes: The ratio of CSM was4.0%(1/25) in GroupA and57.4%(35/61) in Group B (P=0.000).2. Primary tumor stage, lymph nodes down-staging, PR and mammographic malignantcalcification were independent predictors of clinic-pathologic shrinkage mode(P<0.05).The nomogram predicting the risk of NCSM showed a good concordance index(0.869) andgood calibration.3. Mode I, II, III, IV, and V was observed in23(37.7%),17(27.9%),5(8.2%),9(14.8%),7(11.5%), and18(29.5%),3(4.9%),13(21.3%),20(32.8%),7(11.5%) casesrespectively by MRI and Pathology (P=0.001). The accuracy(AC), sensitivity(SE) andspecificity(SP) predictive value of MRI for pathological shrinkage modes were86.2%,65.6%and91.4%, respectively. The CSM and NCSM were observed in36(59.0%) and25(41.0%) cases by pathology,38(62.3%) and23(37.7%) cases by MRI. Two methods had ahigh consistency in clinic-pathologic shrinkage mode(κ=0.863, P=0.000). The AC, SE andSP predictive value of MRI for clinic-pathologic shrinkage modes were91.0%,64.0%and94.8%, respectively. There was not a statistic difference in prediction for CSM and NCSM by MRI(P>0.05). ROC analysis resulted in an AUC of0.928(P=0.000) for MRI to predictthe clinic-pathologic shrinkage mode.4. The longest diameter, maximum cross-section area and volume of the residualtumors measured by MRI and pathology was highly correlated (rs1=0.942, P=0.00;rs2=0.941, P=0.00; rs3=0.903, P=0.00). MRI appears to underestimate pathology in thelongest diameter, maximum cross-section area, but slightly overestimate in volume, andtwo methods had a good consistence(MDs1=0.3cm,95%CI1=-1.43cm~1.9cm, MDs2=1.39cm2,95%CI2=-9.55cm2~12.34cm2, MDs3=-0.433cm3,95%CI3=-7.065cm3~6.199cm3).Conclusions:1. The main shrinkage mode was NCSM after half cycles of NAC, indicating that halfcycles of NAC was not enough to downstage the primary tumor for breast conservingtherapy. Whole cycles of NAC significantly improved the pCR and CSM rate comparedwith half cycles of NAC. The pathologic reconstruction of breast residual tumors couldfully reveal the shrinkage modes of primary breast tumor in women with breast cancer afterNAC. PMSS and3D reconstruction of pathology provided a new platform in this area.2. Based on the clinic-pathological findings of primary tumor, a nomogram to predictshrinkage modes after NAC in breast carcinoma in patients is constructed. The statisticaltool is helpful for individually selecting the patients who can be treated with breast-conserving therapy after NAC.3.3D MRI reconstruction after NAC could simulate real spatial position and sizeofthe residual tumors, and contribute to select patients who received breast-conservingtherapy after NAC with tumor downstaging. |
PDF Full Text Request