Artificial Intelligence-based Radiomics in the Era of Immuno-oncology.

The recent, rapid advances in immuno-oncology have revolutionized cancer treatment and spurred further research into tumor biology. Yet, cancer patients respond variably to immunotherapy despite mounting evidence to support its efficacy. Current methods for predicting immunotherapy response are unreliable, as these tests cannot fully account for tumor heterogeneity and microenvironment. An improved method for predicting response to immunotherapy is needed. Recent studies have proposed radiomics-the process of converting medical images into quantitative data (features) that can be processed using machine learning algorithms to identify complex patterns and trends-for predicting response to immunotherapy. Because patients undergo numerous imaging procedures throughout the course of the disease, there exists a wealth of radiological imaging data available for training radiomics models. And because radiomic features reflect cancer biology, such as tumor heterogeneity and microenvironment, these models have enormous potential to predict immunotherapy response more accurately than current methods. Models trained on preexisting biomarkers and/or clinical outcomes have demonstrated potential to improve patient stratification and treatment outcomes. In this review, we discuss current applications of radiomics in oncology, followed by a discussion on recent studies that use radiomics to predict immunotherapy response and toxicity.

Overcoming acquired resistance to PD-1 inhibitor with the addition of metformin in small cell lung cancer (SCLC).

Metformin has been widely used as the treatment of type II diabetes mellitus for its anti-hyperglycemic effect. In recent years, it has also been extensively studied for its anti-cancer effect as it diminishes immune exhaustion of CD8 + tumor-infiltrating lymphocytes (TILs). It decreases apoptosis of CD8 + TILs, thereby enhancing T cell-mediated immune response to tumor cells. Here, we present a unique case of a patient with small cell lung cancer (SCLC) who exhibited an overall partial response as per Response Evaluation Criteria in Solid Tumors, version 1.1 (RECIST 1.1) since starting metformin in combination with nivolumab therapy. Our patient had been treated with nivolumab monotherapy for 2 years until she had progression of disease. After she was started on metformin along with nivolumab therapy, she has shown a significant durable response for over 6 months. Many patients develop resistance to immunotherapy such as antibodies against cytotoxic T lymphocyte-associated protein 4 (CTLA-4), programmed cell death 1 (PD-1), and programmed cell death ligand 1 (PD-L1). Tumor hypoxia is one of the resistance factors. Signals activated by hypoxic environments in tumors are associated with decreased sensitivity to the PD-1 blockade. Metformin inhibits oxygen consumption in tumor cells in vitro and in vivo, reducing intratumoral hypoxia. These data suggest that metformin can improve susceptibility to anti-PD-1 treatment. To the best of our knowledge, our case is the first reported example demonstrating a proof-of-concept that metformin can contribute to overcoming acquired resistance to PD-1 inhibitors.

Prediction of pathologic complete response using image-guided biopsy after neoadjuvant chemotherapy in breast cancer patients selected based on MRI findings: a prospective feasibility trial.

PURPOSE: Accurate prediction of pathologic complete response (pCR) in breast cancer using magnetic resonance imaging (MRI) and ultrasound (US)-guided biopsy may aid in selecting patients who forego surgery for breast cancer. We evaluated the accuracy of US-guided biopsy aided by MRI in predicting pCR in the breast after neoadjuvant chemotherapy (NAC). METHODS: After completion of NAC, 40 patients with near pCR (either tumor size ≤ 0.5 cm or lesion-to-background signal enhancement ratio (L-to-B SER) ≤ 1.6 on MRI) and no diffused residual microcalcifications were prospectively enrolled at a single institution. US-guided multiple core needle biopsy (CNB) or vacuum-assisted biopsy (VAB) of the tumor bed, followed by standard surgical excision, was performed. Matched biopsy and surgical specimens were compared to assess pCR. The negative predictive value (NPV), accuracy, and false-negative rate (FNR) were analyzed. RESULTS: pCR was confirmed in 27 (67.5%) surgical specimens. Preoperative biopsy had an NPV, accuracy, and FNR of 87.1%, 90.0%, and 30.8%, respectively. NPV for hormone receptor-negative and hormone receptor-positive tumors were 83.3% and 100%, respectively. Obtaining at least 5 biopsy cores based on tumor size ≤ 0.5 cm and an L-to-B SER of ≤ 1.6 on MRI (27 patients) resulted in 100% NPV and accuracy. No differences in accuracy were noted between CNB and VAB (90% vs. 90%). CONCLUSIONS: Investigation using stringent MRI criteria and ultrasound-guided biopsy could accurately predict patients with pCR after NAC. A larger prospective clinical trial evaluating the clinical safety of breast surgery omission after NAC in selected patients will be conducted based on these findings.

Dynamic Contrast-enhanced Breast MRI for Evaluating Residual Tumor Size after Neoadjuvant Chemotherapy.

Purpose To investigate the accuracy of dynamic contrast material-enhanced (DCE) breast MRI for determining residual tumor size after neoadjuvant chemotherapy (NAC). Materials and Methods For this retrospective study, 487 consecutive women (mean age, 47.0 years ± 10.3 [standard deviation]; range, 24-78 years) underwent preoperative DCE MRI following NAC and subsequent surgeries between 2008 and 2011. Tumor size was measured at early-phase, conventional delayed-phase, and late delayed-phase MRI (90, 360, and 590 seconds after contrast material injection, respectively). At histopathologic examination, total tumor size (both invasive and in situ) and the size of invasive tumor alone were separately recorded. Absolute agreement between tumor size at MRI and histopathologic examination was assessed by using intraclass correlation coefficient (ICC) analysis. Factors affecting size discrepancy were assessed by using multiple linear regression analysis. Results Compared with tumor size at histopathologic examination, total tumor sizes showed higher agreement at conventional delayed-phase MRI than at early-phase MRI (ICC, 0.76 vs 0.56; P ˂ .001) and comparable agreement at conventional and late delayed-phase MRI (ICC, 0.76 vs 0.74; P = .55). Lobular histologic features and tumor subtype were independently associated with greater size discrepancy (P ˂ .001). Lobular cancers were underestimated in size compared with ductal cancers (mean size discrepancy, -2.8 cm ± 3.2 vs -0.3 cm ± 1.8; P = .004). Estrogen receptor-positive/human epidermal growth factor receptor 2 (HER2)-negative cancers were underestimated compared with HER2-positive cancers (-0.8 cm ± 2.0 vs -0.3 cm ± 1.7, P = .006) and triple-negative cancers (-0.8 cm ± 2.0 vs 0.3 cm ± 1.7, P ˂ .001). Conclusion Delayed-phase MRI is more accurate than early-phase MRI for evaluating residual breast tumor size after neoadjuvant chemotherapy. Lobular or estrogen receptor-positive/human epidermal growth factor receptor 2-negative cancers are underestimated in size at MRI compared with ductal or other subtypes. © RSNA, 2018 Online supplemental material is available for this article.

Integrated 18F-FDG PET/MRI in breast cancer: early prediction of response to neoadjuvant chemotherapy.

PURPOSE: To explore whether integrated 18F-FDG PET/MRI can be used to predict pathological response to neoadjuvant chemotherapy (NAC) in patients with breast cancer. METHODS: Between November 2014 and April 2016, 26 patients with breast cancer who had received NAC and subsequent surgery were prospectively enrolled. Each patient underwent 18F-FDG PET/MRI examination before and after the first cycle of NAC. Qualitative MRI parameters, including morphological descriptors and the presence of peritumoral oedema were assessed. Quantitatively, PET parameters, including maximum standardized uptake value, metabolic tumour volume and total lesion glycolysis (TLG), and MRI parameters, including washout proportion and signal enhancement ratio (SER), were measured. The performance of the imaging parameters singly and in combination in predicting a pathological incomplete response (non-pCR) was assessed. RESULTS: Of the 26 patients, 7 (26.9%) exhibited a pathological complete response (pCR), and 19 (73.1%) exhibited a non-pCR. No significant differences were found between the pCR and non-pCR groups in the qualitative MRI parameters. The mean percentage reductions in TLG30% on PET and SER on MRI were significantly greater in the pCR group than in the non-pCR group (TLG30% -64.8 ± 15.5% vs. -25.4 ± 48.7%, P = 0.005; SER -34.6 ± 19.7% vs. -8.7 ± 29.0%, P = 0.040). The area under the receiver operating characteristic curve for the percentage change in TLG30% (0.789, 95% CI 0.614 to 0.965) was similar to that for the percentage change in SER (0.789, 95% CI 0.552 to 1.000; P = 1.000).The specificity of TLG30% in predicting pCR) was 100% (7/7) and that of SER was 71.4% (5/7). The sensitivity of TLG30% in predicting non-pCR was 63.2% (12/19) and that of SER was 84.2% (16/19). When the combined TLG30% and SER criterion was applied, sensitivity was 100% (19/19), and specificity was 71.4% (5/7). CONCLUSION: 18F-FDG PET/MRI can be used to predict non-pCR after the first cycle of NAC in patients with breast cancer and has the potential to improve sensitivity by the addition of MRI parameters to the PET parameters.

Pretreatment MR Imaging Features of Triple-Negative Breast Cancer: Association with Response to Neoadjuvant Chemotherapy and Recurrence-Free Survival.

Purpose To investigate whether pretreatment breast magnetic resonance (MR) imaging features are associated with pathologic complete response (PCR) and recurrence-free survival after neoadjuvant chemotherapy (NAC) in patients with triple-negative breast cancer. Materials and Methods Identified were 132 patients with primary triple-negative breast cancers who underwent NAC and pretreatment MR imaging between 2004 and 2010. Three breast radiologists independently reviewed the MR images based on the 2013 Breast Imaging Reporting and Data System lexicon. Presence of intratumoral high signal intensity and peritumoral edema on T2-weighted images was also evaluated. Association of PCR and recurrence-free survival with MR imaging features was assessed by using logistic regression and Cox regression. Bonferroni correction was applied to the P values. Results Among 132 patients, 18 (14%) underwent PCR. Round or oval masses (odds ratio, 3.5 [95% confidence interval: 1.3, 9.7]; P = .02), the absence of intratumoral T2 high signal intensity (odds ratio, 3.8 [95% confidence interval: 1.3, 11.0]; P = .01), and the absence of peritumoral edema (odds ratio, 3.4 [95% confidence interval: 1.2, 9.5]; P = .02) were associated with PCR, but not significantly. After 54 months of median follow-up, there were 41 (31% [41 of 132]) breast cancer recurrences. Peritumoral edema was the only significant variable associated with worse recurrence-free survival (hazard ratio, 4.9 [95% confidence interval: 1.9, 12.6]; P = .001). Conclusion Pretreatment MR imaging features may be associated with PCR and recurrence-free survival in patients with triple-negative breast cancer. © RSNA, 2016 Online supplemental material is available for this article.

Early Stage Triple-Negative Breast Cancer: Imaging and Clinical-Pathologic Factors Associated with Recurrence.

PURPOSE: To determine the imaging and clinical-pathologic factors associated with recurrence in patients with early stage triple-negative breast cancer. MATERIALS AND METHODS: This study was approved by the institutional review board, and the requirement to obtain informed consent was waived. The authors evaluated 398 patients with stage I or II triple-negative breast cancer (median age, 48 years; range, 21-81 years) who were treated between January 2003 and December 2008. Data collected included preoperative breast magnetic resonance (MR) images, mammographic density, patient age, symptoms, family history of breast cancer, histologic tumor characteristics, tumor grade, tumor size, lymphovascular invasion, lymph node involvement, surgery type, margin status, and adjuvant treatment received. Multivariate analysis was performed by using a Cox proportional hazards model, and recurrence-free survival was estimated with the adjusted Kaplan-Meier method. RESULTS: Of the 398 patients, 63 (15.8%) had recurrent disease after a median follow-up of 6.1 years. The absence of preoperative MR imaging (hazard ratio [HR] with multivariate analysis = 2.66; 95% confidence interval = 1.49, 4.75; P < .001), dense breast tissue (HR = 2.77; 95% confidence interval = 1.39, 5.51; P = .004), family history of breast cancer (HR = 2.32; 95% confidence interval = 1.10, 4.90; P = .028), and lymphovascular invasion (HR = 1.83; 95% confidence interval = 1.11, 3.03; P = .019) were found to be independently associated with recurrence. These same factors were also found to be associated with recurrence-free survival. CONCLUSION: The absence of preoperative MR imaging and the presence of dense breast tissue at mammography were associated with an increased risk of recurrence in patients with triple-negative breast cancer.

Early prediction of response to neoadjuvant chemotherapy in breast cancer patients: comparison of single-voxel (1)H-magnetic resonance spectroscopy and (18)F-fluorodeoxyglucose positron emission tomography.

OBJECTIVES: To prospectively compare performances of single-voxel proton magnetic resonance spectroscopy ((1)H-MRS) and (18)F-fluorodeoxyglucose positron emission tomography (FDG-PET) in predicting pathologic response to neoadjuvant chemotherapy (NAC) in breast cancer patients. METHODS: Thirty-five breast cancer patients who received NAC and subsequent surgery were prospectively enrolled. MRS and FDG-PET were performed before and after the 1st NAC cycle. Percentage changes of total choline-containing compounds (tCho) via MRS, and maximum and peak standardized uptake values (SUVmax, SUVpeak) and total lesion glycolysis (TLG) via FDG-PET were measured, and their performances in predicting pathologic complete response (pCR) were compared. RESULTS: Of the 35 patients, 6 showed pCR and 29 showed non-pCR. Mean % reductions of tCho, SUVmax, SUVpeak, and TLG of the pCR group were larger than those of the non-pCR group (-80.3 ± 13.9 % vs. -32.1 ± 49.4 %, P = 0.025; -54.7 ± 22.1 % vs. -26.3 ± 33.7 %, P = 0.058; -60.7 ± 18.3 % vs. -32.3 ± 23.3 %, P = 0.009; -89.5 ± 8.5 % vs. -52.6 ± 36.2 %, P = 0.020). Diagnostic accuracy (area under ROC curve; Az, 0.911) of the % reduction of tCho was comparable to those of %SUVmax (0.822), SUVpeak (0.862), and TLG (0.879) in distinguishing pCR from non-pCR (all P > 0.05). CONCLUSION: MRS showed comparable performance to FDG-PET in early prediction of pCR in breast cancer patients. KEY POINTS: • MRS can predict response to NAC in breast cancer post-1 (st) cycle. • Changes in tCho and SUV after NAC reflect tumour cellularity changes. • MRS can be an alternative to FDG-PET in predicting response to NAC.

Heterogeneity of triple-negative breast cancer: mammographic, US, and MR imaging features according to androgen receptor expression.

OBJECTIVE: Our aim was to determine whether triple-negative breast cancers (TNBCs) with and without androgen receptor (AR) expression have distinguishing imaging features on mammography, breast ultrasound (US), and magnetic resonance (MR) imaging. METHODS: AR expression was assessed immunohistochemically in 125 patients with TNBC from a consecutive series of 1,086 operable invasive breast cancers. Two experienced radiologists blinded to clinicopathological findings reviewed all imaging studies in consensus using the BI-RADS lexicon. The imaging and pathological features of 33 AR-positive TNBCs were compared with those of 92 AR-negative TNBCs. RESULTS: The presence of mammographic calcifications with or without a mass (p < 0.001), non-mass enhancement on MR imaging (p < 0.001), and masses with irregular shape or spiculated margins on US (p < 0.001 and p = 0.002) and MR imaging (p = 0.001 and p < 0.001) were significantly associated with AR-positive TNBC. Compared with AR-negative TNBC, AR-positive TNBC was more likely to have a ductal carcinoma in situ component (59.8% vs. 90.9%, p = 0.001) and low Ki-67 expression (30.4% vs. 51.5%, p = 0.030). CONCLUSION: AR-positive and AR-negative TNBCs have different imaging features, and certain imaging findings can be useful to predict AR status in TNBC. KEY POINTS: • Triple-negative breast cancers have distinguishing imaging features according to AR expression. • AR-positive TNBC is associated with calcifications, spiculated masses, and non-mass enhancement. • Multimodality imaging can help predict androgen receptor status in TNBC.

Magnetic resonance imaging findings of mucinous borderline ovarian tumors: comparison of intestinal and endocervical subtypes.

PURPOSE: The purpose of this study is to compare the MRI features of intestinal and endocervical mucinous borderline ovarian tumors (MBOT). METHODS: Fifty seven and 17 patients with histologically proven intestinal (n = 62) and endocervical (n = 22) MBOT, respectively, underwent preoperative MRI which were reviewed by two radiologists blinded to histology. An array of MRI features and clinical factors (age, cancer antigen 125 [CA-125]) were compared between intestinal and endocervical subtypes using the t test and Chi-square test. Univariate and multivariate logistic regression analyses were performed to evaluate for significant predictors of subtype. RESULTS: There was no significant difference in patient age of intestinal and endocervical MBOT (P = 0.423). CA-125 levels were higher in endocervical MBOT (P = 0.022). Regarding MR features, intestinal MBOT was larger, had more septations, more frequently demonstrated honeycomb loculi, and signal intensity discrepancy while endocervical MBOT was more frequently bilateral with papillary projections (P < 0.05). At multivariate analysis, higher CA-125 (odds ratio [OR] 1.015, P = 0.034) and the presence of papillary projections (OR 11.441, P = 0.024) were the only independent predictive factors of endocervical MBOT. CONCLUSION: Intestinal and endocervical subtypes of MBOT demonstrated significantly different features on MRI. The presence of papillary projection was the only independent MRI feature predictive of endocervical MBOT.

Quantitative MRI morphology of invasive breast cancer: correlation with immunohistochemical biomarkers and subtypes.

BACKGROUND: Breast cancer is a heterogeneous disease with intrinsic molecular subtypes. The different biology and histology of breast cancer exhibit different tumor morphology at breast magnetic resonance imaging (MRI). However, few studies have examined the quantitative relationship between the MRI morphological and immunohistochemical features in breast cancer. PURPOSE: To investigate the correlations between tumor roundness, as quantitatively assessed with MRI and biomarkers or subtypes of breast cancer. MATERIAL AND METHODS: A total of 280 women (mean age, 51 years; range, 28-79 years) with 282 invasive breast cancers (<5 cm) were included. The associations between the tumor roundness (1-100%), as measured using MRI software, and immunohistochemical (e.g. estrogen receptor [ER], progesterone receptor [PR], human epidermal growth factor receptor 2 [HER2], and Ki67) features were evaluated using Pearson's or Spearman's rank correlation coefficients and multiple linear regression analysis. RESULTS: An inverse correlation was observed between the ER (r = -0.408, P < 0.001) or PR (r = -0.248, P < 0.001) scores and tumor roundness, whereas a positive correlation was observed between the Ki67 index and tumor roundness (r = 0.354, P < 0.001). In multiple linear regression, the ER score (P < 0.001) and Ki67 index (P = 0.003) were independent factors determining tumor roundness. Triple-negative tumors (ER, PR, and HER2 negative) showed the highest mean roundness scores compared with the other subtypes (e.g. 67.3% for triple-negative, relative to 55.9% for HER2-enriched, 53.8% for luminal B, and 51.7% for luminal A, P < 0.001). CONCLUSION: Our results suggest that breast tumors with lower ER expression and higher cellular proliferation or biologically aggressive triple-negative tumors are likely to manifest with relatively benign morphologic features.

Breast cancer: early prediction of response to neoadjuvant chemotherapy using parametric response maps for MR imaging.

PURPOSE: To prospectively compare the performance of dynamic contrast material-enhanced (DCE) magnetic resonance (MR) imaging using parametric response map (PRM) analysis with that using pharmacokinetic parameters (transfer constant [K(trans)], rate constant [kep ], and relative extravascular extracellular space [ve]) in the early prediction of pathologic responses to neoadjuvant chemotherapy (NAC) in breast cancer patients. MATERIALS AND METHODS: The institutional review board approved this study; informed consent was obtained. Between August 2010 and December 2012, 48 women (mean age, 46.4 years; range, 29-65 years) with breast cancer were enrolled and treated with an anthracycline-taxane regimen. DCE MR imaging was performed before and after the first cycle of chemotherapy, and the pathologic response was assessed after surgery. Tumor size and volume, PRM characteristics, and pharmacokinetic parameters (K(trans), kep, and ve) on MR images were assessed and compared according to the pathologic responses by using the Fisher exact test or the independent-sample t test. RESULTS: Six of 48 (12%) patients showed pathologic complete response (CR) (pCR) and 42 (88%) showed nonpathologic CR (npCR). Thirty-eight (79%) patients showed a good response (Miller-Payne score of 3, 4, or 5), and 10 (21%) showed a minor response (Miller-Payne score of 1 or 2). The mean proportion of voxels with increased signal intensity (PRMSI+) in the pCR or good response group was significantly lower than that in the npCR or minor response group (14.0% ± 6.5 vs 40.7% ± 27.2, P < .001; 34.3% ± 26.4 vs 52.8% ± 24.9, P = .041). Area under the receiver operating characteristic curve for PRMSI+ in the pCR group was 0.770 (95% confidence interval: 0.626, 0.879), and that for the good response group was 0.716 (95% confidence interval: 0.567, 0.837). No difference in tumor size, tumor volume, or pharmacokinetic parameters was found between groups. CONCLUSION: PRM analysis of DCE MR images may enable the early identification of the pathologic response to NAC after the first cycle of chemotherapy, whereas pharmacokinetic parameters (K(trans), kep, and ve) do not.

Cannabinoid receptor activation inhibits cell cycle progression by modulating 14-3-3ß.

Cannabinoids display various pharmacological activities, including tumor regression, anti-inflammatory and neuroprotective effects. To investigate the molecular mechanisms underlying the pharmacological effects of cannabinoids, we used a yeast two-hybrid system to screen a mouse brain cDNA library for proteins interacting with type 1 cannabinoid receptor (CB1R). Using the intracellular loop 3 of CB1R as bait, we identified 14-3-3ß as an interacting partner of CB1R and confirmed their interaction using affinity-binding assays. 14-3-3ß has been reported to induce a cell cycle delay at the G2/M phase. We tested the effects of cannabinoids on cell cycle progression in HeLa cells synchronized using a double-thymidine block-and-release protocol and found an increase in the population of G2/M phase cells. We further found that CB1R activation augmented the interaction of 14-3-3ß with Wee1 and Cdc25B, and promoted phosphorylation of Cdc2 at Tyr-15. These results suggest that cannabinoids induce cell cycle delay at the G2/M phase by activating 14-3-3ß.

Survival outcomes of breast cancer patients who receive neoadjuvant chemotherapy: association with dynamic contrast-enhanced MR imaging with computer-aided evaluation.

PURPOSE: To retrospectively evaluate whether dynamic contrast agent-enhanced (DCE) magnetic resonance (MR) imaging parameters assessed by a computer-aided evaluation program are associated with recurrence-free and overall survival in breast cancer patients who received neoadjuvant chemotherapy (NAC). MATERIALS AND METHODS: This study was institutional review board approved and informed consent was waived. Between January 2007 and December 2009, 187 consecutive women (mean age, 46.6 years; range, 24-78 years) who had undergone NAC, DCE MR imaging before and after NAC, and surgery for invasive breast cancers (mean size, 5.0 cm; range, 2.0-14.8 cm on surgical histologic analysis) were identified. The tumor size, volume, and kinetic parameters (persistent, plateau, or washout components) were measured with a computer-aided evaluation program on DCE MR images before and after NAC, and their percentage changes were calculated. The Cox proportional hazards model was used to determine the association between DCE MR imaging parameters and recurrence-free survival and overall survival after controlling for clinical-pathologic variables. RESULTS: There were 50 events, including 38 recurrences (29 distant, six local, and three both) and 12 deaths, at a mean follow-up of 47.4 months. At multivariate analysis, a smaller reduction in tumor volume (recurrence-free survival hazard ratio, 5.75; 95% confidence interval: 1.14, 8.64; and overall survival hazard ratio, 2.12; 95% confidence interval: 1.08, 5.69) and a smaller reduction in washout component (recurrence-free survival hazard ratio, 1.15; 95% CI: 1.06, 1.55; and overall survival hazard ratio, 1.26; 95% confidence interval: 1.03, 1.52) after NAC were independent significant variables for worse recurrence-free survival and overall survival. CONCLUSION: Smaller reduction in tumor volume and a smaller reduction in washout component on DCE MR images assessed with computer-aided evaluation after NAC were independent parameters of worse recurrence-free survival and overall survival in breast cancer patients who received NAC.

Breast-conserving surgery after tumor downstaging by neoadjuvant chemotherapy is oncologically safe for stage III breast cancer patients.

BACKGROUND: We conducted a retrospective study to evaluate the local recurrence (LR) rate depending on the use of neoadjuvant chemotherapy (NCT) and to determine the oncologic safety of breast-conserving surgery (BCS) after NCT by comparing LR between patients treated with BCS and mastectomy in clinical stage III breast cancer patients. PATIENTS AND METHODS: Between 2004 and 2007, 166 patients underwent BCS or mastectomy after NCT (NCT group) and 193 patients underwent surgery first (surgery group) in clinical stage III breast cancer patients. Patients whose tumor size became ≤4 cm after NCT, 57 patients underwent mastectomy (mastectomy group) 39 patients underwent preplanned BCS (preplanned BCS group), and 33 patients underwent downstaged BCS (downstaged BCS group). The recurrence rates between the groups and risk factors for LR were analyzed. RESULTS: The 5-year LR-free survival rates were 93.6 % in the NCT group and 95.9 % in the surgery group (P = 0.108). In the NCT group, the 5-year LR-free survival rates were 96.3 % in the mastectomy group, 94.7 % in the preplanned BCS group and 90.9 % in the downstaged BCS group (P = 0.669). High expression of Ki-67 was a predictor of LR in patients in three groups (Hazard ratio 8.300, P = 0.049). CONCLUSIONS: Our findings suggest that BCS after NCT in clinical stage III patients is oncologically safe in terms of LR if breast tumor size is ≤4 cm after NCT and Ki-67 is a predictor of LR after NCT.

Association of tumour stiffness on sonoelastography with axillary nodal status in T1 breast carcinoma patients.

OBJECTIVES: To evaluate whether tumour stiffness on sonoelastography is associated with axillary nodal metastasis in T1 breast carcinoma patients. METHODS: Between May 2006 and December 2010, 200 consecutive women (mean age, 51.6; range, 27 - 81 years) who underwent B-mode ultrasound (US), sonoelastography, and curative surgery with axillary nodal evaluation for clinically node negative T1 breast carcinomas (mean invasive tumour size, 12.4; range, 3 - 20 mm at pathology) were identified. The association between the elasticity score of the tumour and histopathological axillary nodal status was evaluated using a logistic regression model after controlling for imaging and clinicopathological variables of the tumour. RESULTS: The overall incidence of axillary nodal metastasis was 15.5 % (31 of 200). Axillary nodal metastasis was significantly more frequent in tumours with elasticity scores ≥4 than in tumours with elasticity scores <4 (21.7 % vs. 4.2 %; P < 0.001). At multivariate analysis, an elasticity score ≥4 [odds ratio (OR), 6.95; P = 0.004], US size >10 mm (OR, 5.98; P = 0.022), and lymphovascular invasion (OR, 10.68; P < 0.001) of tumours were independently associated with axillary nodal metastasis. CONCLUSIONS: Tumour stiffness on sonoelastography is independently associated with axillary nodal metastasis in T1 breast carcinoma patients. KEY POINTS: • Prediction of axillary nodal status using imaging techniques is valuable. • High ultrasound elasticity scores of T1 tumours were associated with axillary metastasis • Node-positive T1 tumours frequently had elasticity scores 4 or 5. • Sonoelastography might render axillary surgery unnecessary in T1 breast carcinoma patients.

Columnar cell lesions without atypia initially diagnosed on breast needle biopsies: is imaging follow-up enough?

OBJECTIVE: The purpose of this study was to evaluate the underestimation rate and predictive factor of underestimation of columnar cell lesions (CCLs) without atypia diagnosed through breast core needle biopsies (CNBs). MATERIALS AND METHODS: From January 2007 through December 2011, 141 CCLs without atypia, including columnar cell change and columnar cell hyperplasia, were diagnosed in 138 women by CNB. Excisional (n = 16) or imaging follow-up (n = 125) findings were available in all cases. On a per-lesion basis, the underestimation rate and predictive factor of underestimation were evaluated. RESULTS: Among the 16 surgically excised lesions, there were two malignancies (one ductal carcinoma in situ and one invasive ductal carcinoma) and one lobular carcinoma in situ. Overall, the pooled underestimation rate of malignancy was 1.4% (2/141). With regard to lesion variables, the mean lesion size was significantly larger in the underestimation group of CCLs (p = 0.007). Fine pleomorphic morphology of microcalcifications (p < 0.001), the distribution of the microcalcifications (p = 0.007), BI-RADS final assessment (p = 0.001), and imaging-pathologic correlation (p < 0.001) were significantly associated with underestimation. Multivariate analysis showed that fine pleomorphic morphology of microcalcifications (p < 0.0001) was an independent predictor of underestimation in 58 lesions with microcalcifications on mammography. CONCLUSION: The overall underestimation rate of malignancy was 1.4%. Imaging follow-up is reasonable for CCLs without atypia at CNB, especially in small lesions with less suspicious imaging findings. Fine pleomorphic microcalcifications and higher BI-RADS category might be helpful in the prediction of underestimation of a high-risk lesion or malignancy.

Patient age and tumor size determine the cancer yield of preoperative bilateral breast MRI in women with ductal carcinoma in situ.

OBJECTIVE: The objective of our study was to determine the cancer yield of preoperative breast MRI in women with newly diagnosed ductal carcinoma in situ (DCIS) and to identify subgroups of DCIS patients for whom MRI is more likely to identify additional cancers. MATERIALS AND METHODS: A database search for patients with DCIS who had undergone breast surgery between 2007 and 2011 was performed. A total of 308 women (median age, 49 years; range, 25-82 years) with DCIS underwent mammography, sonography, and a preoperative evaluation with MRI. The cancer yield and positive biopsy rate were determined. Patient age at diagnosis, menopausal status, family history of breast cancer, mammographic breast density, index cancer size, nuclear grade, and estrogen receptor (ER) status were examined, and a multivariate logistic regression analysis was performed. RESULTS: The cancer yield was 8% (24/308), and the positive biopsy rate was 53% (24/45). Of the 24 additional cancers, 14 (58%) represented multifocal disease, two (8%) represented multicentric disease, and eight (33%) represented contralateral cancer. In a multivariate analysis, age and index cancer size were factors significantly associated with additional MRI-detected cancers. Compared with women≥50 years, women<50 years had a 3.2-fold higher frequency of detection (95% CI, 1.2-8.9; p=0.02). Compared with women with DCIS lesions smaller than 2.5 cm, women with DCIS lesions 2.5 cm or larger showed a 2.6-fold higher frequency of detection (95% CI, 1.1-6.3; p=0.04). CONCLUSION: Preoperative breast MRI can be considered in patients with DCIS who are <50 years old, have index cancers≥2.5 cm, or fulfill both criteria.

Development of a population pharmacokinetic model of pyrazinamide to guide personalized therapy: impacts of geriatric and diabetes mellitus on clearance.

Objectives: This study was performed to develop a population pharmacokinetic model of pyrazinamide for Korean tuberculosis (TB) patients and to explore and identify the influence of demographic and clinical factors, especially geriatric diabetes mellitus (DM), on the pharmacokinetics (PK) of pyrazinamide (PZA). Methods: PZA concentrations at random post-dose points, demographic characteristics, and clinical information were collected in a multicenter prospective TB cohort study from 18 hospitals in Korea. Data obtained from 610 TB patients were divided into training and test datasets at a 4:1 ratio. A population PK model was developed using a nonlinear mixed-effects method. Results: A one-compartment model with allometric scaling for body size effect adequately described the PK of PZA. Geriatric patients with DM (age >70 years) were identified as a significant covariate, increasing the apparent clearance of PZA by 30% (geriatric patients with DM: 5.73 L/h; others: 4.50 L/h), thereby decreasing the area under the concentration-time curve from 0 to 24 h by a similar degree compared with other patients (geriatric patients with DM: 99.87 µg h/mL; others: 132.3 µg h/mL). Our model was externally evaluated using the test set and provided better predictive performance compared with the previously published model. Conclusion: The established population PK model sufficiently described the PK of PZA in Korean TB patients. Our model will be useful in therapeutic drug monitoring to provide dose optimization of PZA, particularly for geriatric patients with DM and TB.

Nomogram for predicting positive resection margins after breast-conserving surgery.

A positive resection margin after breast-conserving surgery (BCS) is the most important risk factor for tumor recurrence. Re-excision after BCS often results in unnecessary wider excisions, or even mastectomies and poor cosmetic results, as well as increased medical costs and patients' anxiety. A nomogram for predicting positive resection margins may allow the surgeon to develop an individualized surgical plan. Data from 1,034 consecutive breast cancer patients with invasive or in situ breast cancer who initially underwent BCS between January 2008 and December 2009 were used to develop a nomogram for predicting positive resection margins. The nomogram was then validated independently using a cohort of 563 patients who underwent breast surgery in 2010. Multivariate logistic regression analysis showed that microcalcifications (OR 1.574, P = 0.034), grade 4 mammographic density (OR 4.515, P = 0.005), >0.5 cm difference in tumor size between magnetic resonance imaging and ultrasonography (OR 10.001, P < 0.0001), ductal carcinoma in situ (DCIS) on needle biopsy (OR 1.575, P = 0.044), and lobular component on needle biopsy (OR 3.985, P = 0.015) were independent predictors of positive resection margins. These significant variables were used to develop a nomogram for predicting positive resection margins after BCS; the AUCs of the study and the validation cohorts were 0.823 [95 % confidence interval (CI), 0.785-0.862] and 0.846 (95 % CI, 0.800-0.892), respectively. Our new nomogram using 5 variables that were determined before surgery to predict positive resection margins should aid the surgeon in developing individualized surgical plans for breast cancer patients who are scheduled for BCS.