F.N. van Erning

Conditional survival for long-term colorectal cancer survivors in the Netherlands: who do best?

AIMnWith the increase in the number of long-term colorectal cancer (CRC) survivors, there is a growing need for subgroup-specific analysis of conditional survival.nnnMETHODSnAll 137,030 stage I-III CRC patients diagnosed in the Netherlands between 1989 and 2008 aged 15-89 years were selected from the Netherlands Cancer Registry. We determined conditional 5-year relative survival rates, according to age, subsite and tumour stage for each additional year survived up to 15 years after diagnosis as well as trends in absolute risks for and distribution of causes of death during follow-up.nnnRESULTSnMinimal excess mortality (conditional 5-year relative survival >95%) was observed 1 year after diagnosis for stage I colon cancer patients, while for rectal cancer patients this was seen after 6 years. For stage II and III CRC, minimal excess mortality was seen 7 years after diagnosis for colon cancer, while for rectal cancer this was 12years. The differences in conditional 5-year relative survival between colon and rectal cancer diminished over time for all patients, except for stage III patients aged 60-89 years. The absolute risk to die from CRC diminished sharply over time and was below 5% after 5 years. The proportion of patients dying from CRC decreased over time after diagnosis while the proportions of patients dying from other cancers, cardiovascular disease and other causes increased.nnnCONCLUSIONnPrognosis for CRC survivors improved with each additional year survived, with the largest improvements in the first years after diagnosis. Quantitative insight into conditional relative survival estimates is useful for caregivers to inform and counsel patients with stage I-III colon and rectal cancer during follow-up.

Timing of adjuvant chemotherapy and its relation to survival among patients with stage III colon cancer

BACKGROUNDnCurrently available data suggest that delaying the start of adjuvant chemotherapy in colon cancer patients has a detrimental effect on survival. We analysed which factors impact on the timing of adjuvant chemotherapy and evaluated the influence on overall survival (OS).nnnPATIENTS AND METHODSnStage III colon cancer patients who underwent resection and received adjuvant chemotherapy between 2008 and 2013 were selected from the Netherlands Cancer Registry. Timing of adjuvant chemotherapy was subdivided into: ⩽ 4, 5-6, 7-8, 9-10, 11-12 and 13-16 weeks post-surgery. Multivariable regressions were performed to assess the influence of several factors on the probability of starting treatment within 8 weeks post-surgery and to evaluate the association of timing of adjuvant chemotherapy with 5-year OS.nnnRESULTSn6620 patients received adjuvant chemotherapy, 14% commenced after 8 weeks. Factors associated with starting treatment after 8 weeks were older age (Odds ratio (OR) 65-74 versus < 65 years 1.3 (95% confidence interval (CI): 1.14-1.58); OR ⩾ 75 versus < 65 years 1.6 (1.25-1.94)), emergency resection (OR 1.8 (1.41-2.32)), anastomotic leakage (OR 8.1 (6.14-10.62)), referral to another hospital for adjuvant chemotherapy (OR 1.9 (1.36-2.57)) and prolonged postoperative hospital admission (OR 4.7 (3.30-6.68)). Starting 5-8 weeks post-surgery showed no decrease in OS compared to initiation within 4 weeks (Hazard ratio (HR) 5-6 weeks 0.9 (0.79-1.11); HR 7-8 weeks 1.1 (0.91-1.30)). However, commencing beyond 8 weeks was associated with decreased OS compared to initiation within 8 weeks (HR 9-10 weeks 1.4 (1.21-1.68); HR 11-12 weeks 1.3 (1.06-1.59); HR 13-16 weeks 1.7 (1.23-2.23)).nnnCONCLUSIONnOur data support initiating adjuvant chemotherapy in stage III colon cancer patients within 8 weeks post-surgery.

No difference between lowest and highest volume hospitals in outcome after colorectal cancer surgery in the southern Netherlands

AIMnTo investigate the quality of surgical colorectal cancer (CRC) care in the southern Netherlands by evaluating differences between the five hospitals with the lowest volume and the five hospitals with the highest volume.nnnMETHODSnPatients who underwent resection for primary CRC diagnosed between 2008 and 2011 in southern Netherlands were included (n = 5655). The five hospitals performing <130 resections/year were classified low volume; the five hospitals performing ≥ 130 resections/year high volume. Differences in surgical approach, circumferential resection margins (CRM), anastomotic leakage and 30-day mortality between hospital volumes were analysed using Chi(2) tests. Expected proportions anastomotic leakage and 30-day mortality were calculated using multivariable logistic regression. Crude 3-year survival was calculated using Kaplan-Meier curves. Cox regression was used to discriminate independent risk factors for death.nnnRESULTSn23% of patients with locally advanced rectal cancer (LARC) diagnosed in a low volume centre was referred to a high volume centre. Patients with colon cancer underwent less laparoscopic surgery and less urgent surgery in low compared to high volume hospitals (10% versus 32%, p < 0.0001, and 8% versus 11%, p = 0.003, respectively). For rectal cancer, rates of abdominoperineal resections versus low anterior resections, and CRM were not associated with hospital volume. Anastomotic leakage, 30-day mortality, and survival did not differ between hospital volumes.nnnCONCLUSIONnIn southern Netherlands, low volume hospitals deliver similar high quality surgical CRC care as high volume hospitals in terms of CRM, anastomotic leakage and survival, also after adjustment for casemix. However, this excludes LARC since a substantial proportion was referred to high volume hospitals.

Adjuvant chemotherapy is not associated with improved survival for all high-risk factors in stage II colon cancer

Adjuvant chemotherapy can be considered in high‐risk stage II colon cancer comprising pT4, poor/undifferentiated grade, vascular invasion, emergency surgery and/or <10 evaluated lymph nodes (LNs). Adjuvant chemotherapy administration and its effect on survival was evaluated for each known risk factor. All patients with high‐risk stage II colon cancer who underwent resection and were diagnosed in the Netherlands between 2008 and 2012 were included. After stratification by risk factor(s) (vascular invasion could not be included), Cox regression was used to discriminate the independent association of adjuvant chemotherapy with the probability of death. Relative survival was used to estimate disease‐specific survival. A total of 4,940 of 10,935 patients with stage II colon cancer were identified as high risk, of whom 790 (16%) patients received adjuvant chemotherapy. Patients with a pT4 received adjuvant chemotherapy more often (37%). Probability of death in pT4 patients receiving chemotherapy was lower compared to non‐recipients (3‐year overall survival 91% vs. 73%, HR 0.43, 95% CI 0.28–0.66). The relative excess risk (RER) of dying was also lower for pT4 patients receiving chemotherapy compared to non‐recipients (3‐year relative survival 94% vs. 85%, RER 0.36, 95% CI 0.17–0.74). For patients with only poor/undifferentiated grade, emergency surgery or <10 LNs evaluated, no association between receipt of adjuvant chemotherapy and survival was observed. In high‐risk stage II colon cancer, adjuvant chemotherapy was associated with higher survival in pT4 only. To prevent unnecessary chemotherapy‐induced toxicity, further refinement of patient subgroups within stage II colon cancer who could benefit from adjuvant chemotherapy seems indicated.

Recurrence-free and overall survival among elderly stage III colon cancer patients treated with CAPOX or capecitabine monotherapy.

The aim of this study is to investigate the effects of CAPOX and capecitabine on recurrence‐free survival (RFS) and overall survival (OS) among elderly stage III colon cancer patients and to evaluate the effect of (non‐)completion. Patients aged ≥70 years who underwent resection only or who were subsequently treated with CAPOX or capecitabine in 10 large non‐academic hospitals were included. RFS and OS were analyzed with Kaplan‐Meier curves and multivariable Cox regression adjusted for patient and tumor characteristics. 982 patients were included: 630 underwent surgery only, 191 received CAPOX and 161 received capecitabine. Five‐year RFS and OS did not differ between capecitabine and CAPOX (RFS: 63% vs. 60% (pu2009=u20090.91), adjusted HRu2009=u20090.99 (95%CI 0.68‐1.44); OS: 66% vs. 66% (pu2009=u20090.76), adjusted HRu2009=u20090.93 (95%CI 0.64–1.34)). After resection only, RFS was 38% and OS 37%. Completion rates were 48% for CAPOX and 68% for capecitabine. Three‐year RFS and OS did not differ between patients who discontinued CAPOX early and patients who completed treatment with CAPOX (RFS: 61% vs. 69% (pu2009=u20090.21), adjusted HRu2009=u20091.42 (95%CI 0.85–2.37); OS: 68% vs. 78% (pu2009=u20090.41), adjusted HRu2009=u20091.17 (95%CI 0.70–1.97)). Three‐year RFS and OS differed between patients who discontinued capecitabine early and patients who completed treatment with capecitabine (RFS: 54% vs. 72% (pu2009=u20090.01), adjusted HRu2009=u20092.07 (95%CI 1.11–3.84); OS: 65% vs. 80% (pu2009=u20090.01), adjusted HRu2009=u20092.00 (95%CI 1.12–3.59)). Receipt of CAPOX or capecitabine is associated with improved RFS and OS. The advantage does not differ by regimen. The addition of oxaliplatin might not be justified in elderly stage III colon cancer patients.

Intensity of adjuvant chemotherapy regimens and grade III-V toxicities among elderly stage III colon cancer patients

PURPOSEnThe aim of this study was to provide insight in the use, intensity and toxicity of therapy with capecitabine and oxaliplatin (CAPOX) and capecitabine monotherapy (CapMono) among elderly stage III colon cancer patients treated in everyday clinical practice.nnnMETHODSnData from the Netherlands Cancer Registry were used. All stage III colon cancer patients aged ≥70 years diagnosed in the southeastern part between 2005 and 2012 and treated with CAPOX or CapMono were included. Differences in completion of all planned cycles, cumulative dosages and toxicity between both regimens were evaluated.nnnRESULTSnOne hundred ninety-three patients received CAPOX and 164 patients received CapMono; 33% (nxa0=xa063) of the patients receiving CAPOX completed all planned cycles of both agents, whereas 55% (nxa0=xa090) of the patients receiving CapMono completed all planned cycles (Pxa0<xa00.0001). The median cumulative dosage capecitabine was lower for patients treated with CAPOX (163,744xa0mg/m(2), interquartile range [IQR] 83,397-202,858xa0mg/m(2)) than for patients treated with CapMono (189,195xa0mg/m(2), IQR 111,667-228,125xa0mg/m(2), Pxa0=xa00.0003); 54% (nxa0=xa0105) of the patients treated with CAPOX developed grade III-V toxicity, whereas 38% (nxa0=xa063) of the patients treated with CapMono developed grade III-V toxicity (Pxa0=xa00.0026). After adjustment for patient and tumour characteristics, CapMono was associated with a lower odds of developing grade III-V toxicity than CAPOX (odds ratio 0.54, 95% confidence interval 0.33-0.89). For patients treated with CAPOX, the most common toxicities were gastrointestinal (29%), haematological (14%), neurological (11%) and other toxicity (13%). For patients treated with CapMono, dermatological (17%), gastrointestinal (13%) and other toxicity (11%) were the most common.nnnCONCLUSIONnCAPOX is associated with significantly more grade III-V toxicities than CapMono, which had a pronounced impact on the cumulative dosage received and completion of all planned cycles. In this light, CapMono seems preferable over CAPOX.

Adequacy of lymph node yield and staging in rectal cancer should not be determined based on a minimum number of lymph nodes evaluated

Dear Editor: With interest we read the paper ‘A minimum yield of twelve lymph nodes in rectal cancer remains valid in the era of neoadjuvant treatment’ by Lykke et al. published on 5 February 2015 [Epub ahead of print]. In this study, the authors found an association between increased lymph node yield and the detection of node positive disease up to a lymph node yield of 12 to 17 nodes for rectal cancer patients (irrespective of neoadjuvant treatment). They reason that this indicates that a yield of at least 12 lymph nodes is needed to ensure node negative disease, and that guidelines on a minimum yield of 12 lymph nodes are valid. Althoughwe agree with the authors that an adequate lymph node evaluation is important for correct staging and subsequent treatment of patients, we do not believe that their conclusion is justified that the adequacy of lymph node yield and staging can be determined on the basis of a certain minimum number of lymph nodes evaluated. In our opinion, certain important aspects were not taken into account in this study. In the last two decades, the number of lymph nodes evaluated has become a surrogate marker for surgical and pathological excellence in colorectal cancer. Quality initiatives aimed at improving nodal yield have been undertaken and have resulted in an increase in the number of lymph nodes evaluated over time, indicating that room for improvement in surgical and pathological practices existed. However, a ceiling may have been reached beyond which further improvement might not be possible. This can be explained by the fact that the number of lymph nodes available for evaluation is not only influenced by the thoroughness of the surgeon and the diligence of the pathologist but also by a patient’s ability to mount an effective immune response to the tumour. A growing body of literature (i.e. studies by Pages and Galon et al. in the New England Journal of Medicine, Science and Journal of Clinical Oncology) is showing that the biological behaviour of the tumour and host, such as immune response, also affects the number of traceable lymph nodes. A smaller number of lymph nodes reflects a diminished immune response and is also associated with less prominent lymphocytic infiltration into the primary tumour. Consequently, it may be unrealistic to only define a lymph node evaluation as adequate in case 12 or more lymph nodes are yielded. Additionally, other recent studies among colon and rectal cancer patients such as the study by Parsons et al. in the Journal of the American Medical Association in 2011 and the study by Van Erning et al. in European Journal of Cancer in 2014 show that despite increases in lymph node yield over time, the proportion of lymph node positivity has remained unchanged, which makes one question the presumed understaging mechanism. Therefore, we believe that adequacy of lymph node yield and staging in rectal cancer should not be determined based on a minimum number of lymph nodes evaluated. Surgeons and pathologists should strive to remove and evaluate all lymph nodes in the resection specimen, while acknowledging that fewer than 12 lymph nodes might be present in some patients. The absolute number of lymph nodes evaluated should not be used as a marker of quality but should rather be regarded as a reflection of immune response. F. N. van Erning (*) :V. E. P. P. Lemmens Department of Research, Netherlands Comprehensive Cancer Organisation (IKNL), Eindhoven, The Netherlands e-mail: f.vanerning@iknl.nl

Treatment and Outcome of Synchronous Colorectal Carcinomas: A Nationwide Study

BackgroundSynchronous colorectal carcinomas (CRC) occur in 1–8% of patients diagnosed with CRC. This study evaluated treatment patterns and patient outcomes in synchronous CRCs compared with solitary CRC patients.MethodsAll patients diagnosed with primary CRC between 2008 and 2013, who underwent elective surgery, were selected from the Netherlands Cancer Registry. Using multivariable regressions, the effects of synchronous CRC were assessed for both short-term outcomes (prolonged postoperative hospital admission, anastomotic leakage, postoperative 30-day mortality, administration of neoadjuvant or adjuvant treatment), and 5-year relative survival (RS).ResultsOf 41,060 CRC patients, 1969 patients (5%) had synchronous CRC. Patients with synchronous CRC were older (mean age 71xa0±xa010.6 vs. 69xa0±xa011.4xa0years), more often male (61 vs. 54%), and diagnosed with more advanced tumour stage (stage III–IV 54 vs. 49%) compared with solitary CRC (all pxa0<xa00.0001). In 50% of the synchronous CRCs, an extended surgery was conducted (nxa0=xa0934). Synchronous CRCs with at least one stage II–III rectal tumour less likely received neoadjuvant (chemo)radiation [78 vs. 86%; adjusted OR 0.6 (0.48–0.84)], and synchronous CRCs with at least one stage III colon tumour less likely received adjuvant chemotherapy [49 vs. 63%; adjusted OR 0.7 (0.55–0.89)]. Synchronous CRCs were independently associated with decreased survival [RS 77 vs. 71%; adjusted RER 1.1 (1.01–1.23)].ConclusionsThe incidence of synchronous CRCs in the Dutch population is 5%. Synchronous CRCs were associated with decreased survival compared with solitary CRC. The results emphasize the importance of identifying synchronous tumours, preferably before surgery to provide optimal treatment.

The association between body mass index and postoperative complications, 30-day mortality and long-term survival in Dutch patients with colorectal cancer

INTRODUCTIONnThis retrospective study aims to examine the association between body mass index (BMI) and serious postoperative complications, 30-day mortality and overall survival in colorectal cancer (CRC) patients.nnnMATERIALS AND METHODSnAll CRC patients diagnosed between 2008 and 2013 in the south-eastern part of the Netherlands were included. Patients were categorized into four BMI groups: underweight (<18.5), normal weight (18.5u202f≥u202fBMI<25), overweight (25u202f≥u202fBMI<30), and obese (≥30).nnnRESULTSnA total of 7371 CRC patients were included (underweight 133 (1.8%); normal weight 2054 (41.4%); overweight 2955 (40.1%); obesity 1229 (16.7%)). Underweight patients were more likely to have postoperative complications (18.8% vs. 11.7%, adjusted OR 1.95, 95% CI 1.08-3.49) and had a worse 30-day mortality (9.8% vs. 3.3%, adjusted OR 4.37, 95% CI 2.03-9.42) compared to normal weight patients. After stratification for stage (stage I-III and stage IV), underweight was associated with a worse overall survival in both groups compared to normal weight (stage I-III: HR 2.06, 95%CI 1.51-2.80; stage IV: HR 1.65, 95% CI 1.11-2.45). Overweight was associated with an improved overall survival compared to normal weight in both stage groups. Only in stage IV patients obesity was associated with a significant better overall survival compared to stage IV normal weight patients.nnnCONCLUSIONnUnderweight CRC patients were more likely to have postoperative complications and a worse 30-day mortality compared to patients in other BMI categories. The underweight population also has a worse long-term survival while overweight CRC patients and obese stage IV CRC patients were associated with an improved survival compared to normal weight patients.

RE: Effects of adjuvant chemotherapy on recurrence, survival and quality of life in stage II colon cancer patients: a 24-month follow-up

Dear editor, With interest, we read the paper ‘Effects of adjuvant chemotherapy on recurrence, survival, and quality of life in stage II colon cancer patients: a 24-month follow-up’ by Lewis et al. [1]. The authors conclude that stage II colon cancer patients who received chemotherapy were at higher risk for cancer recurrence and all-cause mortality after 24 months compared to those who did not receive chemotherapy. We have some concerns regarding the study methodology and authors’ conclusions. As the authors mention, adjuvant chemotherapy is not the standard of care for all stage II colon cancer patients. It is only considered for patients with high-risk prognostic features. These features include inadequate lymph node sampling, a T4 lesion, perforation or obstruction at presentation, vascular invasion, and/or a poorly differentiated or undifferentiated histology [2]. Therefore, it seems reasonable to assume that in the study by Lewis et al., the patients who received adjuvant chemotherapy exhibited less favourable characteristics than the patients who did not receive adjuvant chemotherapy. The presence of a T4 lesion and vascular invasion has been shown particularly to be associated with a poor prognosis [3, 4]. Unfortunately, no comparison of tumour characteristics in the two groups was provided, which is essential for a good interpretation of the results. Additionally, adjustment for these features is of paramount importance to be able to investigate the (independent) effect of adjuvant chemotherapy on recurrence and survival. The higher recurrence and mortality rates in the group of patients treated with adjuvant chemotherapy could just be the result of the poor prognostic features. Moreover, the results presented by Lewis and colleagues are possibly influenced by selection bias. A substantial portion of eligible patients was not included in the study (i.e. 37 %) and no information on patient and tumour characteristics was provided for this patient group. It seems possible that a more vulnerable group of patients did not participate in the study; that is, patients who had a complicated postoperative recovery or those who had comorbid conditions. In addition, the high proportion of patients treated with adjuvant chemotherapy (i.e. 58 %) in the study population is suspicious. As adjuvant chemotherapy is not the standard care in stage II patients and previous population-based studies reported rates of 16–22 % [3, 5], it is especially likely that patients who did not receive adjuvant chemotherapy did not participate. This would indicate that the patients included in the study and who were not treated with adjuvant chemotherapy represent a selection of fitter patients than in the general population of patients not treated with adjuvant chemotherapy. With regard to the analyses on recurrence, we think the presented results are largely biased. Based on the flowchart the authors provide, it is shown that after 24 months, 261 patients were included. As information on recurrence was collected via telephone interviews, we assume that these data are only available for the 261 patients who completed follow-up. This would mean that for a large part of the patients (i.e. 64 % of eligible patients and 42% of patients enrolled), information on recurrence was lacking. However, the authors seem to base their analyses regarding recurrence on the full enrolled sample * F. N. van Erning f.vanerning@iknl.nl