Search

Background There is an ongoing need to identify pathologic prognosticators in early-stage oral cavity squamous cell carcinoma (OCSCC) to aid selection of patients who may benefit from adjuvant treatment. The objective of this study was to evaluate the prognostic ability of worst pattern of invasion 5 (WPOI-5) defined by the presence of satellite nodules, extratumoural perineural invasion (PNI) and/or extratumoural lymphovascular space invasion (LVI) in low-stage, node negative OCSCC. Methods Retrospective study of 160 patients with T1/T2N0 tumours staged using TNM7 treated surgically. Histologyof the primary tumour was re-reviewed as appropriate to assess for the presence of WPOI-5 parameters. Univariate and multivariate analysis assessing impact of pathological features on survival outcomes was performed Results On univariate analysis WPOI-5 and it’s 3 constituent components of satellite nodules, extratumoural PNI and extratumoural LVI were all significantly associated with disease-specific survival (DSS) and overall survival (OS). On multivariate analysis satellite nodules (odds ratio, 3.58, 95% CI 1.34, 9.55, p=0.01) and extratumoural LVI (odds ratio 10.94, CI 2.22, 53.79, p=0.003) were independently associated with OS. Postoperative radiotherapy was also significantly associated with OS on multivariate analysis (odds ratio 0.42, CI 0.19, 0.89, p=0.02). Conclusion Satellite nodules and extratumoural LVI correlated significantly with survival outcomes in our early-stage OSCC cohort. Further study is required to investigate the benefit of adjuvant treatment in these cases and to ascertain if worst pattern of invasion-5 parameters should be mandatory reporting data elements.

PDF file - 9585K, Supplementary Figure S3: (A) Representative FOXP3 stained TMA (x1.25). (B) Invasive carcinoma showing absence of FOXP3+ lymphocytes (x20). (C-D) Rare intratumoral FOXP3+ lymphocytes (10) (x20, x40, respectively). (G) Representative T-BET stained TMA (x1.25). (H) Invasive carcinoma showing absence of T-BET+ lymphocytes (x20). (I-J) Rare intratumoral T-BET+ lymphocytes (10) (x20, x40, respectively).

PDF file - 7284K, Supplementary Figure S1: (A) Representative CXCL10 stained TMA (x1.25). (B) Invasive carcinoma showing absence of tumoral CXCL10 expression (x20). (C) Low tumoral CXCL10 expression (x20). (D-E) Moderate and high tumoral CXCL10 expression (x20). (F) Representative CXCR3 stained TMA (x1.25). (G) Invasive carcinoma showing absence of tumoral CXCR3 expression. (H-I) Tumoral CXCR3 expression (x20).

PDF file - 10654K, Supplementary Figure S2: (A) Representative CD4 stained TMA (x1.25). (B) Invasive carcinoma showing absence of intratumoral CD4+ lymphocytes (x20). (C-D) Intratumoral CD4+ lymphocytes (x20, x40, respectively). (E-F) Peritumoral CD4+ lymphocytes (x20, x40, respectively). (G) Representative CD8 stained TMA (x1.25). (H) Invasive carcinoma showing absence of intratumoral CD8+ lymphocytes (x20). (I-J) Rare CD8+ intratumoral lymphocytes (10) (x20, x40, respectively). (M-N) Predominantly peritumoral CD8+ lymphocytes (>10) (x20, x40, respectively).

PDF file - 104K, Table S1. Summary of antibodies and conditions of use Table S2. Pathologic Characteristics of the Breast Tumors Table S3. Association between CXCL10 and CD4 and CD8 peritumoral expression Table S4. (A) Association between FOXP3 and Molecular Biomarkers Table S4. (B) Association between T-BET and Molecular Biomarkers Table S5. (A) Association of FOXP3 and subgroups Table S5. (B) Association of T-BET and subgroups Table S6. (A) Association between FOXP3 and Clinicopathologic parameters Table S6. (B) Association between T-BET and Clinicopathologic parameters

Purpose: Breast carcinomas, including basal and hereditary cases, often present with a prominent tumoral lymphocytic infiltrate. Chemokines could play a role in attracting these cells and contribute to tumor progression. We explored tumoral expression of CXCL10 and determined the relationship between CXCL10 and lymphocytic infiltrate in a cohort of breast cancers.Experimental Design: Using tissue microarrays of 364 breast tumors, we evaluated expression of CXCL10 and its receptor, CXCR3, in relation to histopathologic features, biomarkers, and lymphocyte markers. In addition, we overexpressed CXCL10 and CXCR3 in MCF7 breast cancer cells and monitored T-lymphocyte migration and invasion.Results: Forty-five percent of tumors expressed CXCL10, and a significant association was found with CXCR3 and lymphocytic infiltrate. Further characterization of the lymphocytic infiltrate revealed an association with CXCL10 expression for peritumoral CD4+ and CD8+ lymphocytes. CD8+ intratumoral lymphocytes, FOXP3+ regulatory T cells (Tregs), and T-BET+ TH1 cells were associated with BRCA1 and basal tumors. Conditioned media from MCF7 cells overexpressing both CXCL10 and CXCR3 increased T-lymphocyte migration and invasion.Conclusions: Our findings suggest that CXCL10 may act in a paracrine manner, affecting the tumor microenvironment, and in an autocrine manner, acting on the tumor cells themselves and may play a role in tumor invasiveness and progression. The CXCL10-CXCR3 axis can serve as a potential target in BRCA1 and basal breast cancers, which present with a prominent lymphocytic infiltrate and a poor prognosis. Clin Cancer Res; 19(2); 336–46. ©2012 AACR.

PDF file - 56K

Supplementary Figure S4: Confirmation of CXCL10 and CXCR3 overexpression in CXCL10-CXCR3 cells. (A) ELISA showing the average level of CXCL10 secreted into the media over 24h by 1x107 CXCL10-CXCR3 or empty vector cells, respectively. (B) Flow cytometry results showing the mean fluorescence intensity of CXCR3 and the percent of CXCR3 positive cells in the CXCL10-CXCR3 and empty vector cell populations, respectively. Filled histograms represent isotype-matched control, and empty histograms represent CXCR3-PE.

Purpose Systematic identification of all 4 parathyroid glands has been recommended during total thyroidectomy (TT); however, it is unclear whether this strategy necessarily translates into optimized functional parathyroid preservation. We wished to investigate the association between number of parathyroids identified intraoperatively during TT, and incidence of incidental parathyroidectomy, and postoperative hypoparathyroidism. Methods Retrospective review of prospectively maintained database of 511 consecutive patients undergoing TT at an academic teaching hospital. The association between number of parathyroid glands identified intraoperatively and incidence of biochemical hypocalcaemia (defined as any calcium n first 48 h after surgery), symptomatic hypocalcaemia; permanent hypoparathyroidism (defined as any hypocalcaemia or need for calcium or vitamin D > 6 months after surgery), and incidental parathyroidectomy, was investigated. The association between number of parathyroid glands visualized and postoperative parathyroid hormone (PTH) levels was investigated in a subset of 454 patients. Results Patients in whom a greater number of parathyroids had been identified had a significantly higher incidence of biochemical and symptomatic hypocalcaemia, and significantly lower postoperative PTH levels, than patients with fewer glands identified. There were no significant differences in incidence of permanent hypoparathyroidism or incidental parathyroidectomy. On multivariate analysis, malignancy, Graves disease, and identification of 3–4 parathyroids were independent predictors of biochemical hypocalcaemia. For symptomatic hypocalcaemia, identification of 2–4 parathyroids, and identification of 3–4 parathyroids, were significant. Conclusions Systematic identification of as many parathyroid glands as possible during TT is not necessary for functional parathyroid preservation.

Pathological margin assessment is an essential component of surgical management of oral cavity squamous cell carcinoma (OCSCC), however, in many studies, variable definitions of involved margins have been used. The purpose of the present study was to compare the prognostic ability of involved margins according to Royal College of Pathologists (RCPath) and College of American Pathologists (CAP) guidance.Retrospective study of 300 patients with previously untreated OCSCC undergoing definitive surgical management. Main specimen margin status was defined according to RCPath guidance and CAP guidance. "Final margin status", incorporated the results of frozen sections and extra tumour bed resections. The prognostic impact of each margin definition was studied using univariate analysis, and in multivariate models including T-stage (AJCC 8th edition), nodal status (pN+), extranodal extension (ENE), and use of adjuvant radiotherapy.Both RCPath and CAP positive margins were associated with local recurrence (LR), disease-specific survival (DSS), and overall survival (OS) on univariate analysis, while final margin status was associated with LR and DSS, but not OS. On multivariate analysis, only CAP positive main specimen margin status was independently associated with LR (odds ratio 2.44, 95% CI 1.37, 4.34), DSS (odds ratio 2.28, 95% CI 1.31, 3.82), and OS (odds ratio 1.59, 95% CI 1.04, 2.42).Involved main specimen margin as defined by CAP guidance has the advantage of being an independent prognosticator of LR and survival in our cohort.

Human papilloma virus (HPV)-positive oropharyngeal squamous cell carcinomas (OPSCC) tend to have good outcomes, however a subset does not share this favourable prognosis. The aim of this paper is to investigate the utility of tumour cell anaplasia and multinucleation as prognostic markers in oropharyngeal squamous cell carcinoma. Retrospective review of 104 patients with OPSCC or squamous cell carcinoma of unknown primary site (SCCUP) who underwent primary resection and/or lymph node dissection. Slides of both primary and nodal metastatic disease were assessed for the presence of anaplasia and multinucleation. 53 patients were HPV-positive. Anaplasia was more frequent in males (p = 0.005), smokers (p = 0.003), and HPV-negative disease (p = 0.04). HPV status and > 10 pack-year smoking history were independent predictors of recurrence-free survival (RFS) and disease-specific survival (DSS). Neither anaplasia, nor multinucleation, at the primary site or in cervical metastases, had any significant impact on RFS or DSS. We did not find either anaplasia or multinucleation to have any significant prognostic impact in OPSCC. However, given the small number of adverse events in the HPV-positive cohort, we may have lacked sufficient power to detect significance in what was the subgroup of primary interest. Our study highlights the challenge of identifying markers of poor prognosis in HPV-positive OPSCC. ELECTRONIC SUPPLEMENTARY MATERIAL: The online version of this article (10.1007/s12105-019-01081-7) contains supplementary material, which is available to authorized users.

Background: Two major changes to the staging of oral cavity squamous cell carcinoma (OCSCC) were adopted in TNM8: (1) depth of invasion is now used for T staging and (2) extranodal extension for N staging. The aim of this study was to evaluate if TNM8 stratifies OCSCC patients more accurately than TNM7 based on overall survival (OS) statistics and hazard discrimination. Methods: Retrospective study of 297 patients with OCSCC who underwent surgery at our institution. Clinical and pathological data were previously populated from review of medical charts and histological reports. Slides were re-reviewed for depth of invasion measurements. Patients were staged using both TNM7 and TNM8 with overall survival statistics analysed. Results: Overall 118 patients (39.7%) were upstaged using TNM8. Both TNM7 and TNM8 stage categories were highly significant for OS (all p values < 0.0001). Hazard discrimination analysis showed that TNM7 could only differentiate stage III from stage IV disease with significance (OS p = 0.01). In comparison TNM8 could distinguish between stage II and III disease (OS p = 0.047) and between stage III and IV disease (OS p = 0.004). Subsite analysis suggested that both editions of the staging system perform best for tongue primaries. Conclusions: Although TNM8 showed improved hazard discrimination in comparison to TNM7, problems with discriminative ability persisted with 8th edition staging criteria. Large scale validation studies will be required to direct future refinement of the staging rules and to establish if the continued use of a single staging system for all oral cavity subsites is appropriate.

BACKROUND: The purpose of this study is to assess the impact of trimodal therapy [surgery, chemotherapy and external beam radiotherapy (EBRT)] in patients with anaplastic thyroid cancer (ATC) treated with curative intent. MATERIALS AND METHODS: Retrospective review of patients with ATC treated at a tertiary referral centre between January 2009 and June 2020. Data were collected regarding demographics, histology, staging, treatment and outcomes. RESULTS: Seven patients (4 female) were identified. Median age was 58 years (range 52–83 years). All patients received EBRT with concurrent doxorubicin. Six patients received surgery followed by chemoradiotherapy (CRT), and one underwent neoadjuvant CRT followed by surgery. Median radiological tumour size was 50mm (range 40–90 mm). Six patients had gross extrathyroidal extension and three had N1b disease. Prescribed radiotherapy schedules were 46.4 Gy in 29 bidaily fractions (n = 2, treated 2010), 60 Gy in 30 daily fractions (n = 2), 66Gy in 30 fractions (n = 2) and 70 Gy in 35 fractions (n = 1; patient received neoadjuvant CRT). CRT was discontinued early for two patients due to toxicities. At median follow up of 5.8 months, 42.9% (3/7) patients were alive and disease-free. Only one patient developed a local failure. Three patients died from distant metastases without locoregional recurrence. CONCLUSIONS: Despite poor prognosis of ATC, selected patients with operable tumours may achieve high locoregional control rates with trimodal therapy, with possibility of long-term survival in select cases.

Objectives: The updated American Society of Clinical Oncology/College of American Pathologists (ASCO/CAP) guidelines (2013) for human epidermal growth factor receptor 2 (HER2) testing in breast cancer recommend repeat testing at excision of HER2-negative grade 3 breast tumors. This study aimed to identify the rate of HER2 discordance in this cohort of cases. Methods: All HER2-negative grade 3 tumors diagnosed at a single institution over a 15-month period had reflex repeat HER2 testing at excision . HER2 testing was performed in accordance with ASCO/CAP guidelines using immunohistochemistry (IHC) and dual in situ hybridization (ISH). Results: One hundred cases were identified over the study period. HER2 was amplified at excision in three cases. The discordant tumors showed equivocal IHC at excision with low-level amplification on dual ISH. All discordant cases showed equivocal IHC on core needle biopsy (CNB) specimens and/or tumor upgrade at excision. Conclusions: Our series demonstrated a high concordance rate (97%) for HER2 at excision in grade 3 breast tumors with a negative core biopsy result. These findings suggest that reflex repeat HER2 testing of all these cases, which has significant cost and workload implications, may not be justified. Features that may indicate HER2 heterogeneity, such as equivocal IHC on CNB specimens or tumor upgrade at excision, may help refine selection of cases for repeat testing.

Fibroadenomas (FA) are the most common benign tumor in the female breast. Most are managed conservatively provided there is clinical, radiologic, and pathologic concordance. However, surgical excision is typically recommended for cellular fibroepithelial lesions or those lesions with clinical, radiologic, or pathologic features concerning for phyllodes tumor (PT). Some studies have suggested surgical excision in all FA >30 mm to reduce core needle biopsy (CNB) sampling errors. The aim of our study was to evaluate, in the absence of any other concerning clinicopathologic features, whether surgical excision of FA was warranted based on size criteria alone. Cork University Hospital is a large academic center in Southern Ireland. Its breast cancer center provides both a screening and symptomatic service and diagnoses approximately 600 cancers per year. The breast histopathological data base was reviewed for all CNBs from January 1, 2010, to June 30, 2015, with a diagnosis of FA that went on to have excision at our institution. We excluded all cellular fibroepithelial lesions and those cases with co-existent lobular neoplasia, ductal carcinoma in situ, invasive carcinoma, atypical ductal hyperplasia, or lesions which would require excision in their own right. Cases in which the radiologic targeted mass was discordant with a diagnosis of FA were also excluded. Patient demographics and preoperative radiologic size and the radiologic target were recorded in each case. All radiology was reviewed by a breast radiologist prior to inclusion in the study, and there was histologic radiologic concordance with a diagnosis of FA in all cases. A total of 12,109 consecutive radiologically guided CNB were performed January 2010-June 2015; 3438 with a diagnosis of FA were identified of which 290 cases went on to have surgical excision. Of those 290 cases; 98.28% (n = 285) were confirmed as FA on excision. The remaining 1.72% (n = 5) had atypical features-FA with LCIS (n = 1), benign PT (n = 3), and invasive ductal carcinoma (n = 1). Our study suggests that, excision based solely on size is not warranted in clinical and radiologically concordant cases with a diagnosis of FA on CNB.

Importance The presence of extracapsular spread (ECS) of metastatic nodes is considered a poor prognosticator in head and neck cancer, with postoperative chemoradiation therapy often recommended over radiation therapy alone in such cases. However, there is less clarity regarding the effect of ECS on human papillomavirus–associated oropharynx squamous cell carcinoma (OPSCC) or carcinoma of unknown primary site (CUP). Objective To investigate the association of ECS according to human papillomavirus status in OPSCC and CUP with survival. Design, Setting, and Participants This investigation was a retrospective cohort study performed between August 1998 and March 2015 at an academic teaching hospital. Participants were 83 patients with OPSCC (n = 62) or CUP (n = 21) undergoing neck dissection as part of initial treatment. Main Outcome and Measures Human papillomavirus status was determined by p16 immunohistochemistry. The presence of ECS was extrapolated from pathology reports, and the extent of ECS was determined by rereview of original pathology slides. Disease-specific survival (DSS) and recurrence-free survival (RFS) were assessed. Results Among 83 patients (71 male), there were 45 p16-positive and 38 p16-negative tumors. Fifty-one patients had ECS, which was graded as extensive in 43 cases. The median follow-up was 31 months for all patients and 50 months for surviving patients. Among the entire cohort, adverse predictors of RFS were p16-negative status (hazard ratio [HR], 9.4; 95% CI, 3.3-27.2) and ECS (HR, 6.5; 95% CI, 2.0-21.6). Adverse predictors of DSS were p16-negative status (HR, 16.8; 95% CI, 3.9-71.2) and ECS (HR, 8.3; 95% CI, 2.0-35.3). Among p16-negative patients, ECS was significantly associated with worse RFS (HR, 9.7; 95% CI, 1.3-72.3) and DSS (HR, 8.7; 95% CI, 1.1-62.7). In contrast, among p16-positive patients, ECS had no effect on RFS (HR, 1.1; 95% CI, 0.2-7.8) or DSS (HR, 1.2; 95% CI, 0.1-18.7). Conclusions and Relevance The presence of ECS appears to be associated with survival in OPSCC and CUP according to p16 status. Our findings raise questions regarding the benefits of postoperative chemoradiation therapy in p16-positive patients with ECS.

Objectives: The purpose of this study was to determine the optimum number of cells that should be counted when scoring human epidermal growth factor receptor 2 (HER2) brightfield dual-color in situ hybridization (BDISH), including cases with HER2 /chromosome 17 (Chr17) ratios in the 1.80 to 2.20 range. Methods: In total, 131 cases of breast carcinoma with HER2 immunohistochemistry and BDISH were included. For cases with a HER2 /Chr17 ratio of less than 1.80 or more than 2.20 (n = 115), BDISH scoring was performed for 60 cells using three tumor fields, and for cases with a HER2 /Chr17 ratio of 1.80 to 2.20 (n = 16), scoring was performed for 120 cells using six tumor fields. Mean HER2 /Chr17 ratio and HER2 copy number were calculated for cumulative cell counts. Results: The HER2 status as determined by the HER2 /Chr17 ratio or HER2 copy number was unchanged following counting of additional cells in 100% of cases with ratio of less than 1.80 or more than 2.20. The HER2 status of two cases with ratios of 1.80 to 2.20 changed from positive to negative following counting of 120 cells. Conclusions: Our findings support recommendations to score 20 nuclei in conjunction with careful assessment of immunohistochemistry and scan of the BDISH slide to identify areas of heterogeneity. Scoring of additional cells/fields is likely not of benefit and might be a disadvantage since the scorer moves out of the area of strongest signal.

Importance The follicular variant (FV) of papillary thyroid carcinoma (PTC) is an important subtype that can be difficult to diagnose using preoperative cytologic analysis. Objective To compare conventional and FV PTC with regard to preoperative cytologic diagnosis using a tiered thyroid cytologic reporting system, tumor size at diagnosis, presence of invasion, and implications on prognostic scores. Design, Setting, and Participants This retrospective study was conducted in an academic teaching hospital and included 99 patients with conventional (n = 65) or FV (n = 34) PTC. Interventions Preoperative thyroid cytologic findings, originally reported using the tiered British Thy system, were recategorized according to the Bethesda classification system. Pathologic features recorded included tumor size, presence of extrathyroid extension (ETE), and metastases. Prognostic scores were calculated according to the MACIS system. Main Outcomes and Measures Differences in patient demographics, preoperative cytologic findings, tumor pathologic features, and prognostic risk categories between conventional and FV PTC were studied. Results There were no differences in patient age or sex. Cytologic findings from FV PTC were significantly more likely to be reported in a lower-risk category than those from conventional PTC for (1) malignant vs lower-risk category (22 [56%] vs 2 [8%]); (2) suspected malignant or malignant vs lower-risk category (26 [66%] vs 6 [23%]); and (3) follicular neoplasm or higher-risk category vs lower-risk category (34 [87%] vs 10 [38%]) ( P P = .03). The mean size of FV PTC lesions (25.9 mm) at the time of pathologic diagnosis was significantly greater than that of conventional PTC lesions (15.5 mm) ( P = .02). Even after exclusion of “coincidental” carcinomas, FV PTC tumors were significantly larger than conventional PTC tumors (31.7 vs 22.4 mm; P = .03). In contrast, FV PTC was significantly less likely to show ETE (0 of 34 vs 10 of 65; P = .01). There were no significant differences between FV PTC and conventional PTC in proportion of patients in intermediate- and high-risk prognostic groups combined (21 [62%] vs 38 [58%]) ( P = .83) or in mean MACIS scores (4.68 and 4.38, respectively; P = .18). Conclusions and Relevance Preoperative cytologic findings from FV PTC were more likely than those from conventional PTC to indicate a lower risk category, and FV PTC tumors were larger at time of diagnosis. On the other hand, owing to a lower incidence of ETE in conventional PTC, there was no difference in prognostic score at diagnosis.

Introduction The role of sentinel lymph node biopsy (SLNB) in ductal carcinoma in situ (DCIS) is controversial. This study evaluates the risk of clinically relevant SLN metastasis following a core needle biopsy (CNB) diagnosis of pure DCIS. Materials and methods Cases that underwent SLNB following a CNB diagnosis of pure DCIS at our institution over a 4.5 year period were evaluated. Parameters including the DCIS characteristics on CNB, the rate of upstaging to invasive carcinoma at excision and the SLNB result were recorded. Results Of 296 patients with a CNB diagnosis DCIS, 181 had SLNB (62%). The rate of invasion at excision in those undergoing SLNB was 30% (54/181). SLN metastasis was detected in 7/181 cases (4%), including 6 cases with isolated tumour cells only (3.5%) and only 1 case with a macro-metastatic deposit (0.5%). Conclusion The risk of clinically significant SLN metastasis following a CNB diagnosis of DCIS is extremely low, despite a relatively high rate of upstaging to invasive carcinoma at excision. Our findings support the opinion that SLNB is not warranted following a CNB diagnosis of DCIS, particularly for those patients undergoing breast conservation surgery.

The objectives of this study were to determine the prognostic significance of subgrouping estrogen receptor (ER)-positive breast tumors into low- and high-risk luminal categories using Ki67 index, TP53, or progesterone receptor (PR) status. The study group comprised 540 patients with lymph node negative, invasive breast carcinoma. Luminal A subtype was defined as being ER positive, HER2 negative, and Ki67 low (

Purpose: Breast carcinomas, including basal and hereditary cases, often present with a prominent tumoral lymphocytic infiltrate. Chemokines could play a role in attracting these cells and contribute to tumor progression. We explored tumoral expression of CXCL10 and determined the relationship between CXCL10 and lymphocytic infiltrate in a cohort of breast cancers. Experimental Design: Using tissue microarrays of 364 breast tumors, we evaluated expression of CXCL10 and its receptor, CXCR3, in relation to histopathologic features, biomarkers, and lymphocyte markers. In addition, we overexpressed CXCL10 and CXCR3 in MCF7 breast cancer cells and monitored T-lymphocyte migration and invasion. Results: Forty-five percent of tumors expressed CXCL10, and a significant association was found with CXCR3 and lymphocytic infiltrate. Further characterization of the lymphocytic infiltrate revealed an association with CXCL10 expression for peritumoral CD4+ and CD8+ lymphocytes. CD8+ intratumoral lymphocytes, FOXP3+ regulatory T cells (Tregs), and T-BET+ TH1 cells were associated with BRCA1 and basal tumors. Conditioned media from MCF7 cells overexpressing both CXCL10 and CXCR3 increased T-lymphocyte migration and invasion. Conclusions: Our findings suggest that CXCL10 may act in a paracrine manner, affecting the tumor microenvironment, and in an autocrine manner, acting on the tumor cells themselves and may play a role in tumor invasiveness and progression. The CXCL10-CXCR3 axis can serve as a potential target in BRCA1 and basal breast cancers, which present with a prominent lymphocytic infiltrate and a poor prognosis. Clin Cancer Res; 19(2); 336–46. ©2012 AACR.