Epidermal growth factor receptor (EGFR) as a predictive and prognostic marker in patients with advanced colorectal cancer (aCRC): The MRC COIN trial experience

359

Background: KRAS mutation has been shown to be a more effective (though negative) biomarker for selection of patients for EGFR targeted therapy in aCRC. However, positive EGFR immunohistochemistry (IHC) remains a license requirement and was an inclusion criterion in most trials to date. The MRC COIN trial recruited 2445 pts into 3 arms of oxaliplatin + fluoropyrimidine +/- cetuximab without prior EGFR assessment. This trial provides a unique opportunity to definitively examine the role of EGFR IHC as prognostic and predictive marker and potentially the evidence required to remove this assessment from the license for this drug.

Methods: Formalin-fixed paraffin embedded (FFPE) tissue was stained retrospectively for EGFR using Dako kit in a national reference lab. Results were assessed by 3 reviewers (BJ, SS, RA) using digital imaging software in a blinded fashion, then by BJ/SS providing consensus for discrepancies. EGFR scoring was assessed as a prognostic variable in association with selected patient, tumor and biochemical data. Cut off points examined for +ve vs -ve tumours, in terms of total tumour cells demonstrating membrane staining, were: 0% vs >0%; <10% vs ≥10%; <20% vs ≥20%.

Results: EGFR IHC was adequately assessed for 1621 pts (66% of randomised), 22% were negative (0%) and 78% positive (>0%), balanced across arms. EGFR was not prognostic for PFS within KRAS wt pts at the standardized cut off point 0% vs >0% HR=1.11 95% CI 0.91-1.36 p=0.31 but was at <10% vs ≥10% (HR=1.27 95% CI 1.07-1.52 p=0.008) this was robust to other prognostic variables. No effect was seen for overall response or survival. There was no prognostic effect for the KRAS mutant group. In the 1065 assessable pts randomised to +/- cetuximab, no evidence of EGFR IHC as a predictive marker for response or survival outcomes was observed for the addition of cetuximab to chemotherapy (OS HR=1.11 95% CI 0.70-1.75 p=0.66; PFS HR=0.95 95% CI 0.64-1.43 p=0.82).

Conclusions: Extensive assessment of samples from this trial suggest a role for EGFR IHC as a prognostic marker in KRAS wt aCRC but refute the predictive value embedded within the licence for cetuximab used in combination with chemo in first-line therapy.

[Table: see text]

Biomarker potential of EGFR gene copy number by FISH in the phase III EXTREME study: Platinum-based CT plus cetuximab in first-line R/M SCCHN

6005

Background: Platinum-based CT + cetuximab is the first systemic therapy in ∼30 years to show a survival benefit vs platinum-based CT in first-line R/M SCCHN (Vermorken JB, et al. N Engl J Med 2008;359:1116–1127). A retrospective analysis has evaluated the influence of EGFR gene copy number, determined by FISH, on clinical outcome in the EXTREME study. Methods: Pts were randomized to 3-weekly cycles of platinum-based CT (cisplatin 100 mg/m2 or carboplatin AUC 5, day 1; 5-fluorouracil 1000 mg/m2/day continuous infusion, days 1–4) with or without cetuximab (initial dose 400 mg/m2, then 250 mg/m2 weekly). The proportion of FISH+ cells per pt (FISH score) was determined using 5 different enrichment models. Tumors were also classified as FISH+ or FISH- using the Colorado scoring system. Results: In the overall population (n=442), addition of cetuximab significantly improved median OS (10.1 vs 7.4 months; p=0.04). No association between FISH score and OS, PFS, or best overall response was determined for any enrichment model. Pts with Colorado FISH+ tumors were evenly distributed between the CT + cetuximab (50/158) and CT-alone (51/154) arms of the FISH- evaluable population (71% of ITT population). Colorado FISH status had no influence on OS in either treatment arm, on PFS in the CT-alone arm, or on RR in the CT + cetuximab arm (see table ). In the CT + cetuximab arm, pts with FISH+ tumors had a lower risk of progression than pts with FISH- tumors. Higher RRs among pts with FISH- tumors in the CT-alone arm may have been due to twice as many nonevaluable response observations in the FISH+ vs the FISH- population (percentage of pts with SD or PD was comparable). Conclusions: EGFR gene copy number, as determined by FISH, is not a predictive biomarker for cetuximab efficacy in R/M SCCHN.

[Table: see text]

[Table: see text]

Cardiogenic shock due to thrombotic thrombocytopenic purpura

We report the case of a 49-year-old man with thrombotic thrombocytopenic purpura (TTP) leading to cardiogenic shock. Laboratory data were typical for TTP with thrombocytopenia and microangiopathic hemolytic anemia. The electrocardiogram recorded significant ST-segment elevations in the anterior and inferior leads. In addition' coronary angiography showed normal epicardial coronary arteries with slow flow. The patient died due to electromechanical dissociation six hours after admission. During autopsy typical features of thrombotic thrombocytopenic purpura were found. Histological preparation of the heart showed a diffuse myocardial necrosis due to microvascular thrombosis. Cardiac involvement is common in TTP but extended myocardial necrosis has been reported in only a few cases.

Comparative genomic hybridization (CGH) analysis of neuroblastomas--an important methodological approach in paediatric tumour pathology

Comparative genomic hybridization (CGH) was applied to 35 neuroblastomas to obtain a global view of genetic imbalances. Results were validated by means of Southern blot hybridization (detection of N-myc amplification), loss of heterozygosity (LOH) studies (detection of deletion 1p), and interphase cytogenetics [dual labelling fluorescence in situ hybridization (FISH) of centromeric 17 and erbB-2]. CGH allowed sensitive detection of N-myc amplification and chromosome 1p deletion, representing the most established prognostic markers of neuroblastoma. In addition, a high rate of chromosome 17 aberrations (63 per cent) with possible prognostic relevance was observed. Previously unreported high level copy number increases indicating oncogene amplification were mapped to chromosome subbands 2p13-14 and 3q24-26. Other recurrent regional chromosomal aberrations were localized on 11q, 12q, 13q, 14q, and 15q. CGH results were fully consistent with data of Southern blot analysis and LOH study, as well as interphase cytogenetics. These results show that CGH is a sensitive method for the detection of all prognostically relevant genetic alterations in neuroblastomas; that CGH considerably simplifies the detection of these alterations, resulting in a single methodological approach; and that CGH is a powerful tool to elucidate previously unknown genetic changes in neuroblastomas.

Pulmonary contusion: CT vs plain radiograms

In experimentally induced pulmonary contusions, CT (n = 27) and chest X-ray (n = 24) findings were compared with the findings at autopsy. Twenty-seven of 27 (100%) pulmonary contusions were visible by CT immediately after trauma compared with 9 of 24 (37.5%) in the chest X-ray. After 30 min follow-up, 18 of 24 (75%) lesions were seen on the plain film. Five of 24 (21%) contusions escaped detection on conventional radiographs. Computed tomography underestimated lesion size in 5 of 60 (8%) measurements, conventional radiographs in 21 of 36 (58%) measurements. Pathological examination never revealed a pulmonary contusion that was not demonstrated by CT. Therefore, pulmonary contusion seems unlikely in a trauma patient with normal pulmonary CT.

[Computed tomography of lung contusion. An experimental study]

A new experimental model has been developed to determine whether computed tomography has any advantages for the investigation of lung contusions. In 27 dogs, chest wall deformation was produced at speeds varying from 17.3 m/sec to 45.7 m/sec, leading to lung contusion. After the injury, 27 out of 27 (100%) CT examinations demonstrated the contusions, but only nine out of 24 (37.5%) were seen on conventional radiographs. In contrast to conventional radiography, which either failed to show the presence of a contusion, or which under-estimated the extent by more than one cm, in more than half the cases CT agreed with the pathological findings in more than 90%. No lung contusions were found at autopsy which had not been demonstrated by CT.