Risk Factors for Recurrent Papillary Thyroid Carcinoma

Objective: To identify risk factors associated with disease recurrence among Filipinos with papillary thyroid carcinoma (PTC). Methods: Design: Retrospective cohort study Setting: Tertiary National University Hospital Participants: 76 patients diagnosed with papillary thyroid carcinoma, classified as low and low-to-intermediate risk (2015 ATA classification) that underwent total thyroidectomy with or without neck dissection from 2010-2014 and were followed up from 10 months to 5 years. Log rank and Cox regression analyses were used to determine significant risk factors for recurrence. Results: 29 (38.15%) had recurrence. On univariate analysis, age, tumor size, multifocality, extrathyroidal extension, presence of lateral neck nodes and RAI therapy were statistically associated with recurrence. However, on multivariate analysis, no clinicopathologic factor was statistically associated with recurrence. Conclusion: Age of >45 years, female sex, tumor size of >2 cm, multifocality, presence of microscopic extrathyroidal extension, and lymph node metastasis might contribute to the recurrence of papillary thyroid cancer while post-operative radioactive ablation may have some protective effect. However, this study suggests that other factors must be included in the model to better understand the relationship between these factors and recurrence. Keywords: papillary thyroid cancer, thyroid neoplasm, recurrence

The impact of prostate cancer (PC) margin extent (ME) at radical prostatectomy (RP) on biochemical relapse-free survival (bRFS).

83 Background: Surgical margin involvement (M+) by PC at RP is associated with suboptimal bRFS; however, the interaction of M+ with coincident high-risk clinicopathologic and treatment factors obscures accurate estimation of recurrence risk. The objective of this study is to determine whether ME permits risk stratification. Methods: Retrospective analysis of clinicopathologic factor association with bRFS. Eligible patients underwent RP alone for clinically localized PC. Patients with metastatic disease, PSA > 30 at diagnosis, pathologic involvement of seminal vesicles or lymph nodes at RP, insufficient follow-up, or receipt of pre-RP or adjuvant therapy were excluded. Slides from RP specimens with close or positive margins were re-reviewed by study pathologists blinded to outcome. Results: From 2002-2010, 667 patients were eligible for analysis. The median age was 61 yrs (range, 43-76), and all had cT1-2 disease (83% T1c), with median PSA 5.6 (0.9-28.0; 85% ≤10). Robot-assisted RP was employed in 141 cases (21%). Two hundred ten patients (31%) had M+, with single maximal ME 3mm (0.1-23), and cumulative ME 4mm (0.1-34). At median follow-up of 102 months (13-184), 149 patients (22%) had recurrence, with estimated 8-year bRFS rates of 85%/56% for M-/+ patients (p < 0.01). Multivariable analysis identified PSA, Gleason score (GS), extraprostatic extension, and M+ as associated with bRFS. Specific to patients with involved margins, the combination of ME and GS permitted recurrence risk stratification, with a low-risk subset identified (GS≤6 and ME < 3mm; Table). Conclusions: The current investigation suggests that GS6 patients with maximal or cumulative ME < 3mm appear to have favorable early 8-year bRFS following RP. GS6 patients with wider ME and GS ≥7 with any extent M+ appear to have suboptimal bRFS. RP pathology reports should include ME details, in order to more precisely estimate risk of subsequent disease recurrence. [Table: see text]

Disease control and survival outcomes for rectal adenocarcinoma (RA) managed with neoadjuvant chemoradiotherapy (nCRT): Performance of serum CEA changes as a biomarker.

552 Background: nCRT is commonly employed in advanced or distal RA, with validated associations between tumor response and disease control; however, the identification of patients likely to respond remains elusive. The present investigation seeks to determine whether absolute CEA levels and changes during treatment are associated with pathologic complete response (pCR), freedom from failure (FFF), disease specific survival (DSS), and overall survival (OS). Methods: Retrospective analysis of clinicopathologic and treatment factor association with outcomes. Eligible patients underwent nCRT followed by mesorectal resection; patients with pre-nCRT evidence of metastasis or who did not undergo resection were excluded. CEA levels were recorded at 3 times: pre-nCRT (C1), post-nCRT/pre-op (C2), and post-op (C3; < 45 days of surgery, or < 225 days if post-op C given). Univariate analysis was performed to identify clinicopathologic factor association with pCR, FFF, DSS, and OS. The absolute and relative changes in interval CEA levels were computed and included as factors in the analysis. Results: From 2003-11, 71 patients were eligible. At median follow-up of 57 months (range, 5-124), 20 patients had RA recurrence and 25 had died (19 recurrent). nCRT resulted in pCR for 19 patients (27%), which was associated with DSS (p = 0.04). The estimated 5-year FFF, DSS, and OS for the entire population were 69%, 77%, and 72%, respectively. Clinical and pathologic N-stage, nodal ratio, ypTNM stage, and stage migration post-CRT were significantly associated with FFF, DSS, and OS. The relative change from C1 to C2 was significantly associated with pCR (exp(b); p = 0.031). Absolute C3 was associated with FFF (1.091; 0.001), DSS (1.183; 0.018), and OS (1.092; 0.002); however, absolute and relative changes at C2 and C3 as compared with C1 did not demonstrate associations with disease control or survival endpoints. Conclusions: Within the present study, changes in CEA levels following CRT were associated with pCR. Additionally, initial post-op CEA demonstrated associations with disease control and survival endpoints.

The 21-gene breast cancer assay in small (<1cm) tumors.

33 Background: Tumor size is a clinicopathologic factor often used in early stage breast cancer (EBC) management to estimate prognosis and make decisions about therapy. While in general, small tumors have a lower chance of recurring than large tumors, some patients with small tumors have poor outcomes. The 21-gene breast cancer assay (OncotypeDX) yields a Recurrence Score result that predicts the 10-yr risk of distant recurrence (DR) and the likelihood of chemotherapy (CT) benefit in patients (pts) with ER+ EBC. This study reviews the clinical evidence and experience of the 21-gene assay in subcentimeter (subCM) tumors and shows that the Recurrence Score is more reflective of the underlying tumor biology. Methods: A meta-analysis of Genomic Health validation and supportive studies reporting Recurrence Score result and tumor size was conducted in order to describe the distribution of Recurrence Score results. Results: There were 10 GHI-sponsored studies (7,094 pts) that presented Recurrence Score results as a function of tumor size. The studies varied in the recorded tumor sizes, ranging from < = 0.5 cm to > 5 cm. Four studies specifically examined the distribution of Recurrence Score results in subCM tumors and identified a total of 629 pts out of 2,912 pts (21.6%) that fit the size criteria of < 1cm. A broad distribution of score results was observed in subCM tumors in the four studies combined: 61% low, 25% int, 15% high. Conclusions: These results demonstrate that not all pts with subCM tumors have low risk disease, and that the 21-gene assay can help inform the risk of DR more precisely as well as predict the likelihood of CT benefit. Approximately 40% of pts had intermediate or high score results, suggesting an increased potential for DR and CT benefit. The clinical implications of this meta-analysis are that size doesn’t tell the whole story. There is underlying biology reflected in the Recurrence Score that can better inform the treatment decision for pts with ER+ EBC. [Table: see text]

Somatic mutations of EGFR and KRAS in pure bronchioloalveolar carcinomas and adenocarcinomas of the lung with bronchioloalveolar features

21111 Background: EGFR mutations are more common in lung cancers from female, non-smoking patients of Asian ethnicity with adenocarcinoma histology. We investigated the relationship between clinicopathologic parameters and mutations of EGFR and KRAS in patients with pure bronchioloalveolar carcinomas (BACs) and adenocarcinomas (ADCs) with BAC features, not previously treated with EGFR inhibitors. Methods: We retrospectively reviewed the files of patients who were diagnosed with pure BACs or ADCs with BAC features between February 2001 and May 2005. Demographics, stage and smoking histories were recorded. Paraffin embedded tissue blocks were retrieved and re-examined using the WHO 1999/2004 histological criteria for lung cancers. EGFR expression, detected by immunohistochemistry, was scored from 0 to +3. Expression was considered positive when the score was +1, or higher. DNA was isolated from paraffin blocks (containing at least 75% tumor cell content) and exons 18–21 of the EGFR gene and exon 2 of the KRAS gene were amplified by PCR and bi-directionally sequenced. The chi-squared test and regression analysis were used to identify parameters correlating with EGFR or KRAS mutations. Results: A total of 46 patients were eligible for this study. Tissue was available from 43 patients, 19 with pure BAC and 24 with ADC with BAC features. There were 27 (62.8%) males and 16 (37.2%) females, all of Greek descent. Mean age was 62.7 years (range, 41–79). There were 34 smokers and former smokers (79%) and 9 never-smokers (21%). Tumors were classified as mucinous (20), non- mucinous (19) or mixed (4). We detected somatic EGFR mutations in 7 patients (including common mutations, delEx19, L858R, G719C), and KRAS codon 12 mutations in 13 patients. EGFR and KRAS mutations were mutually exclusive. No clinicopathologic parameters correlated with the presence of EGFR or KRAS mutations. Conclusions: Among Greek lung cancer patients with pure BACs or ADCs with BAC features, unselected for response to EGFR inhibitors, the incidence of KRAS or EGFR mutations was 30.2% and 16.3% respectively. EGFR and KRAS mutations were mutually exclusive but no clinicopathologic factor predicted the presence of mutations in either gene. Further molecular analysis is ongoing. No significant financial relationships to disclose.

Late relapse associated with CEP 17 polysomic breast cancer

20011 Background: The clinicopathologic features associated with chromosome enumeration probe 17 (CEP 17) polysomy by fluorescence in situ hybridization (FISH) are not well defined. CEP 17 polysomy is frequently encountered in assessing Her-2neu amplification, which has become an important adjuvant therapeutic target. One mechanism to increase Her-2neu gene copy is polysomy. In this analysis the prognostic and predictive factors associated with CEP 17 polysomy are compared to similar factors in Her-2neu gene amplification. Methods: All cases of Her-2neu gene amplification and CEP 17 polysomic breast cancers from 2000 to present were abstracted. The polysomy group was defined as at least 3 gene copies and was further subdivided into two groups; 3–4 copies and ≥ 5 copies. This resulted in 193 cases of invasive breast cancer, which became the study cohort. Polysomic status and her-2neu gene amplification was compared to age, estrogen receptor (ER), progesterone receptor (PR), grade, tumor size, cell type, mitotic rate, nodal status, number of positive nodes and relapse. Descriptive statistics were used for categorical variables. The χ2 test was used to compare each clinicopathologic factor. Results: Both the ER and PR status was significantly different in the 3 groups. For the Her-2neu amplified, the polysomic 3–4, and the polysomic ≥ 5, the ER positive status was 53%, 67% and 81%, respectively (p = .012). Histologic grade, tumor size and nodal status were not significantly different between groups. Lobular pathology was present in 15% of the polysomy ≥ 5 group, 8% of the polysomy 3–4 group and 1.7% of the amplified group and this difference was significant (p = .03). Relapse disease status was significantly more frequent in the polysomy ≥ 5 group (18.5%) compared to the amplified group (2.6%) (p = .007). Within the relapsed group the median time to relapse was 4 years for the amplified patients versus 12 years and 10 years for the polysomic patients. Conclusion: The incidence of CEP 17 polysomy varies in the literature from 10–50% depending on definitions. Our study is unique in that we divided the polysomic group into 3–4 copies which can be complicated by proliferative activity versus ≥ 5 copies. A significant association between relapse and polysomic breast cancer is described in our dataset. No significant financial relationships to disclose.