A paradox explained? Patients with delayed diagnosis of symptomatic colorectal cancer have good prognosis.

OBJECTIVE

To investigate the impact on outcome of delay between referral and diagnosis in colorectal cancer (CRC).

PATIENTS AND METHODS

One hundred and fifty-four patients were studied after excluding from a consecutive series of 411 with CRC, those with factors known to affect the prognosis that may also have affected the speed of diagnosis. These were advanced disease, emergency admission or surgery, referral with diagnosis already made, and tumours treated by colonoscopic polypectomy alone. Possible causative factors were compared between early and late diagnosis groups. For assessment of symptom risk, the Department of Health criteria were used.

RESULTS

Forty-four patients had Referral to Diagnosis Interval (RDI) > or = 50 days ('Late'), and 110 had RDI < 50 days ('Early'). In the Late group there were only 2 deaths from cancer and 93.7% cancer-specific five year survival (c5ys), compared with 22 and 65.3%, respectively, in the Early one (P = 0.007). There were more Duke's A cases in the Late group (38.6%vs 15.2%, P = 0.006), but this did not fully explain the improved survival. Comparisons for each Duke's Stage showed improved c5ys for Late Duke's B ones (100% of 16 vs 60.3% of 54, P = 0.039). Late patients had more low risk symptoms than Early ones, both overall (31.8%vs 13.7%, P = 0.013) and in Duke's B cases (56%vs 15.3%, P = 0.003). Tumours were smaller in the Late group (length 35.3 vs 41.6 mm, P= 0.04); this difference was confined to the Duke's A patients and sigmoid tumours. Late sigmoid tumours were not only shorter (32.4 vs 45.9 mm, P = 0.02) but also were all cured (c5ys 100% of 18 vs 60.3% of 23, P = 0.011). There were no differences between Late and Early groups in: age (mean 69.9 years), sex (male 57.7%), date of diagnosis (mean December 1998), ASA comorbidity index (mean 1.9), number of lymph nodes found in the operative specimen (mean 8.6), or histological grading (moderate differentiation 94.4%).

CONCLUSION

In the context of modern rapid access clinics, symptomatic CRC patients with delay between referral and diagnosis (even if this is several months or occasionally more than a year) have less aggressive tumours and markedly better long-term cure rate than their earlier diagnosed counterparts. Attempts to speed up further the diagnosis would be a waste of time and resources, being unlikely to make an appreciable difference to the overall cure rate.