Consecutive MCRC patients not eligible, due to comorbidities and/or age, for expanded clinical program or ongoing phase II trial proposing intensive regimens consisting of triplet chemotherapy plus targeted agent, were treated in clinical practice with first line medical and/or surgical treatments, chosen among those in indication for MCRC treatment and approved by Agenzia Italiana del Farmaco (AIFA) for administration in label in Italian public hospitals, and published in Gazzetta Ufficiale Repubblica Italiana (‘Elenco dei Medicinali erogabili a totale carico del Servizio sanitaria nazionale’, Gazzetta Ufficiale Repubblica Italiana N.1, 2 Gennaio 2009). Thus, it was not a clinical trial and approval by ethics committee and institutional review board was not necessary, because patients were treated with conventional treatments without any additional medical intervention out of the best common clinical practice. Patients had histological confirmed diagnosis of MCRC, age ≥18 years, PS ≤2. Criteria to define patients unfit, or not eligible for intensive regimens were: age ≥75 years; uncontrolled severe diseases; cardiovascular disease (uncontrolled hypertension, uncontrolled arrhythmia, ischemic cardiac diseases in the last year); thromboembolic disease, coagulopathy, preexisting bleeding diatheses; proteinuria >1 g/24 h. Patients were classified according to Cumulative Illness Rating Scale (CIRS) (10). Treatment options were tailored according to age (< or ≥75 years), patient’s fitness (PS, CIRS), KRAS genotype. Patients with PS 3 were not treated. All patients provided written, informed consent to the proposed in label treatment option.

Medical treatments included triplet, doublet, or mono-chemotherapy. Triplet FIr/FOx schedule consisted of weekly timed-flat-infusion 5-FU (TFI 5-FU), associated to weekly alternating CPT-11 or L-OXP (41): TFI/5-FU (Fluorouracil Teva; Teva Italia, Milan, Italy), 750–900 mg/m²/die, over 12 h (from 10:00 pm to 10:00 am), days 1–2, 8–9, 15–16, 22–23; CPT-11 (Campto; Pfizer, Latina, Italy), 120–160 mg/m², days 1 and 15; l-OXP (Eloxatin; Sanofi-Aventis, Milan, Italy), 70–80 mg/m², days 8 and 22; cycles every 4 weeks. Other triplet, doublet and mono-regimens were administered according to previously reported schedules (7,41,42). Targeted agents were: BEV (Avastin; Roche, Welwyn Garden City, UK), 5 mg/kg, days 1 and 15; cetuximab (Erbitux; Merck, Darmstadt, Germany), 400 mg/m² initial dose, then 250 mg/m²/week.

Genetic analyses were performed on paraffin-embedded tissue blocks from primary tumor and/or metastatic sites, as previously reported (4). Genotype status was assessed for KRAS codon 12 and 13 mutations by direct sequencing. KRAS exon 2 sequence was performed from PCR-amplified tumor DNA using the Big Dye V3.1 Terminator kit, electrophoresis in ABI PRISM 3130xl Genetic Analyzer, and analysis using the GeneMapper Analysis software version 4.0 (Applied Biosystems, Foster City, CA, USA).

Activity, efficacy, and prognostic relevance of first line treatments, and KRAS genotype on clinical outcomes were evaluated. Patients were classified according to: metastatic extension, L-L and O/MM (3,4); age, non-elderly (<65 years), young-elderly (≥65 <75 years), old-elderly (≥75 years); CIRS stage primary, intermediate, secondary. Clinical evaluation of response was made by CT scan; PET was added based on investigators’ assessment. Follow-up was scheduled every two-three months up to disease progression or death. L-L patients were evaluated at baseline and every two-three cycles of treatment by a multidisciplinary team, to evaluate resectability defined according to reported categories (3). Liver metastasectomies were defined as R0, if radical surgery, R1, if radioablation was added. Surgery was recommended >4 weeks after BEV discontinuation.

Clinical criteria of activity and efficacy were ORR, resection rate of metastases, PFS and OS: ORR, evaluated according to RECIST criteria (43); pathologic complete response, defined as no residual cancer cells in surgical specimens; PFS and OS, evaluated using the Kaplan-Meier method (44). PFS was defined as the length of time from the beginning of treatment and disease progression or death (resulting from any cause) or to the last contact; OS as the length of time between the beginning of treatment and death or to last contact. Log-rank test was used to compare PFS and OS according to medical treatment, KRAS genotype, metastatic extension, age and comorbidity stage (45).

Forty patients unfit for intensive regimens, among 72 consecutive MCRC (56%), were treated with (Table I): medical treatments, 37 patients (92.5%); surgery, 3 (7.5%). First line medical treatments: triplet, 18 (45%); doublet, 15 (37.5%); mono-therapy, 4 (10%). Among 39 KRAS evaluated patients (97.5%), 23 (59%) were wild-type and 16 (41%) mutant. Clinical features of the 37 patients who underwent first line medical treatments were (Table IIA): male/female ratio, 22/15; median age, 75 years; young- and old-elderly, 28 (76%) and 20 (54%), respectively; PS 0, 15 (41%) and 1–2, 22 (59%); metastatic disease metachronous 24%, synchronous 76%. Liver metastases, 26 patients (70%); L-L 8 (22%), O/MM 29 79%). Distribution of patients according to age and comorbidity stage (Table IIB): non-elderly 9 (24%), young-elderly 8 (22%), old-elderly 20 (54%); CIRS stage primary 4 (11%), intermediate 15 (40%), secondary 18 (42%). KRAS mutations detected in 15 patients were: codon 12, 13 (36.1%), specifically c.35 G>A (G12D), 7 (19.4%), c.35 G>T (G12V), 6 (16.6%); codon 13, 2 (5.5%), c.37 G>T (G13V), 1 (2.7%) and c.38 G>A (G13D), 1 (2.7%).

Medical treatments were tailored according to age and CIRS stage. Triplet regimens were administered in 18 patients (49%): non-elderly 6, young-elderly 4, old-elderly 8; CIRS primary 2, intermediate 10, secondary 6. Doublet regimens were administered in 15 patients (40%): non-elderly 3, young-elderly 3, old-elderly 9; CIRS primary 1, intermediate 4, secondary 10. Mono-regimens were administered in 4 patients (11%): young-elderly 1, old-elderly 3; CIRS primary 1, intermediate 1, secondary 2.

Among the 37 patients who underwent medical treatments, 10 were not evaluable for activity: 7 (19%) did not receive at least 2 cycles of treatment; 3 were on-treatment. The intent-to-treat analysis of 27 patients showed ORR 37% (α 0.05, CI ± 19) (Table IIIA). We observed 10 objective responses: 9 partial (33%) and 1 complete (CR 4%); 9 stable diseases (33%); 8 progressive diseases (30%). Disease control rate was 67% (α 0.05, CI ± 18). After median follow-up of 8 months, median PFS was 7 months (1-13+): 28 events occurred. Median OS was 13 months (1+−23+): 22 events occurred (Fig. 1A). R0 liver metastasectomies were performed in 3 patients (8%): 3 out of 8 L-L (37.5%). No surgery-related complications were reported. Overall, 1 clinical plus 1 pathologic CR were reported (7%); 1 patient showed a progressive disease at 8 months; 1 patient was progression-free at 10 months. Pathologic CR was obtained in 1 KRAS wild-type patient (33%), with primary rectal tumor and a single L-L metastasis. Twelve patients (32%) received, at least, a second line treatment.

Among 7 evaluable L-L patients, ORR was 71%; 3 performed liver metastasectomies (43%) and 1 cCR (14%); median PFS 11 months (3–13+ months); median OS 12 months (3–13+ months). Among 20 evaluable O/MM patients, ORR was 25%; median PFS 6 months (1–12 months); median OS 13 months (1+−23+ months). Clinical outcome (PFS and OS) in L-L compared to O/MM patients was not significantly different (Fig. 1B).

Among 16 evaluable patients treated with triplet regimens (Table IIIB), ORR was 37.5% (α 0.05, CI ± 24). We observed 6 partial responses (37.5%); 5 stable diseases (31%); 5 progressive diseases (31%). Median PFS was 8 months (3–12): 14 events occurred. Median OS was 12 months (3–23+ months): 12 events occurred. Secondary metastasectomy was performed in 1 patient (6%). Among 15 patients treated with doublet regimens (Table IIIB), ORR was 44% (α 0.05, CI ± 34). We observed 3 partial responses (33%); 1 CR (11%); 3 stable diseases (33%); 2 progressive diseases (22%). Median PFS was 8 months (1–13+): 9 events occurred. Median OS was 15 months (1+−23+ months): 7 events occurred. Among 4 patients treated with mono-regimens, median PFS was 5 months (3–6 months), median OS 6 months (3−13+ months). Among 3 patients who underwent surgery as first line treatment, median PFS was not reached (3+−19+ months); median OS not reached (3+−19+ months). PFS and OS were not significantly different in patients treated with triplet compared to other first line treatments (p=0.947 and 0.557, respectively), and to doublet regimens (p= 0.885 and 0.616, respectively) (Fig. 2).

Moreover, PFS and OS were not significantly different in non-elderly and young-elderly compared to old-elderly patients (p=0.240 and 0.750, respectively), and in primary and intermediate CIRS stage compared to secondary stage patients (p=0.494 and 0.364, respectively).

Among 14 KRAS wild-type patients evaluable for activity, ORR was 50% (α 0.05, CI ± 27) (Table IIIA). We observed 7 objective responses: 6 partial (43%) and 1 CR (7%); 4 stable diseases (29%); 3 progressive diseases (21%). Disease control rate was 79% (α 0.05, CI ± 22). Liver metastasectomies were performed in 3 patients (14%), 3 out of 7 L-L (43%). Median PFS was 8 months (1+−13+ months), 15 events occurred (71%). Median OS was 13 months (1+−23+ months), 11 events occurred. Among 12 KRAS mutant patients evaluable for activity, ORR was 25% (α 0.05, CI ± 26). We observed 3 partial responses (25%); 5 stable diseases (42%); 4 progressive diseases (33%). Disease control rate was 67% (α 0.05, CI ± 28). No liver metastasectomies were performed. Median PFS was 6 months (1–11 months), 12 events occurred (80%). Median OS was 8 months (3–18 months), 10 events occurred. KRAS wild-type compared with mutant patients showed significantly different PFS (p=0.043), but not OS (Fig. 1C). KRAS c.35 G>A mutant patients showed significantly worse PFS and OS compared to wild-type (p=0.000, and 0.049, respectively) (Fig. 3A and B), and to other mutant patients (p=0.020 and 0.048, respectively) (Fig. 3C and D). No different clinical outcomes were reported in other than c.35 G>A KRAS mutant compared to wild-type patients (Fig. 3E and F). PFS and OS were also significantly worse in c.35 G>A KRAS mutant patients compared to other mutant plus wild-type patients (p=0.000, and 0.021, respectively) (Fig. 3G and H).