Based on the entry criteria of this evaluation (see above), data were extracted from 561/661 records. Files of 554 out of 630 stage I/II patients and of 93 stage III patients within a median follow-up time of 4.1 and 1.5 years, respectively, were analysed. The detection mode of recurrent disease could be categorised in 127 stage I/II patients presenting first recurrences, and in all 73 stage III patients developing second recurrences.
In 561 out of 661 patients with primary melanoma an initial staging was performed. Imaging procedures () detected synchronous metastases in 31 out of 561 patients. Twenty-seven out of 31 patients were upstaged to IIIA/B disease (4.7%), only four patients (0.6%) were shown to be stage IV with asymptomatic distant metastases. After excision of the primary tumour in 630 stage I/II melanoma patients, a total of 127 (20%) FR were identified over time. Distribution of Breslow thickness and recurrence rate among stage I/II patients in our study group correlated well with other study cohorts as shown in . Five hundred and fifty-four out of 630 patients participated in the follow-up programme for more than 6 months. Ninety-five out of 127 first relapses were detected in the follow-up of patients in stage I/II with 88 recurrences (90.7%) being noted in the first 5 years and with 75 (77.3%) of the metastases detected by the end of third year. Eight out of 197 (4.1%) patients with a ‘low risk’ melanoma (pT
mm) relapsed; only two after the fifth year. The metastatic pattern of first relapse is shown in and was similar to other reports (Kersey et al, 1985
; Fusi et al, 1993
; Soong et al, 1998
). Ninety-three patients with surgically resected loco-regional metastases (24 initial stage IIIA/B and 69 former stage I/II patients) were enrolled into the follow-up programme, in 60 out of 93 stage III patients a relapse was documented (64.5%) within a median time of 7.8 months.
Number of documented initial staging tests performed at the time of primary diagnosis.
First recurrence (FR) rates in stage I/II patients by Breslow thickness
Figure 2 Distribution pattern of 127 first recurrences in 630 FU I/II patients. The recurrence pattern was classified into local (satellites or in-transit metastases), regional (regional lymph nodes), distant (viscera; distant (sub-)cutis or lymph nodes) or combinations (more ...) Efficacy of staging methods
The detection rate including the number of false-positive results of each staging method separated by the clinical phase (initial staging, follow-up stage I/II and III) is given in . At initial staging, 2554 imaging procedures were performed in 561 patients yielding 31 metastases (true-positive) and 202 false-positive results leading to further technical examinations. In follow-up of stage I/II patients 30 metastases (32%) were detected by the patient and triggering a premature visit, however, 45 of the remaining 65 metastases detected at this stage were detected by the doctor (). At any phase of melanoma staging and follow-up patient' history and physical examination was the most successful diagnostic tool indicating the vast majority of all relapses (around 70%) in patients attending the follow-up program. Although lymph node sonography was the best performing method among the imaging procedures, detection rate was substantially lower (between 15–20%) compared to physical examination and detection rate of lymph node sonography notably decreased at the different phases of melanoma disease. Chest X-ray and sonography of the abdomen showed extremely low detection rates (below 10%) when used for routine follow-up in stage I/II and stage III patients (). Overall, the detection of 65 out of 95 first (68.5%) and 41 out of 60 second relapses (68.3%) could be attributed to the scheduled follow-up activities. Thirty out of 127 first relapses (24%) developed in patients not enrolled into the follow-up programme at the time of diagnosis. At initial staging, 325 scintigraphies of the bones and 282 cranial CT scans were performed without revealing any metastases whereas lymph node sonography was remarkably effective (16% detection rate (five out 31 metastases detected)) ().
Efficacy of diagnostic methods at initial staging and in follow-up of stage I/II and stage III melanoma patients
Cost-efficiency of staging methods
Absolute and relative expenses (including costs caused by false-positive results) of each staging method were summarised (). Absolute costs for clinical assessment were the highest in all clinical phases making up to 53% of total costs for follow up. Detection rates were high (). Costs for physical assessment ranged between €1100 (at initial staging) to €7300 (follow-up in stage I/II) per detected metastasis. Sonography of the lymph nodes proved to be the most cost consuming technical screening method with about 25% of total expenses at each phase of follow-up. Costs ranged between €4400 at initial staging to €13
300 (follow-up of stage I/II) per detected metastasis. In contrast, total costs for screening by chest X-ray ranged between 17.2 to 9.2% (), however, detection rates were low () and costs to detect a metastasis ranged between €2800 (in stage III) to €13
500 at initial staging. Among initial staging methods, scintigraphy of bones and cranial CT-scans were most expensive and comprised 67.4% of total costs. Considering ‘detection rate’ and ‘relative costs’, the efficiency ratio for each method clearly demonstrated the physical examination to be superior to all other staging at all follow-up phases. The costs for each relapse detected within each risk category of primary tumour during follow-up in stage I/II () varied from €35
mm) to €2326. (pT>4.0
mm). Due to the low rate of relapses in patients enrolled into follow-up program in stage I/II in 5th and following years, the mean price per detected recurrence drastically increased from €5806 (±€1.289, s.d.; years 1–4) to €18
558 (±€6.706, s.d.; years 5–8).
Summarised cost (€) caused by each screening methods at the time of initial staging and during routine follow-up of stage I/II+III patients
Distribution of costs (€) for follow-up of localised melanoma by Breslow thickness and costs per detected first recurrence (FR)
Survival after relapse was analysed by mode of detection of relapse. There was no difference in survival between patients with symptomatic relapse (84%) and asymptomatic relapse (16%; ). Also the small subset of patients during follow-up in stage I/II in whom a first recurrence was detected by lymph node sonography (n=9) did not survive significantly longer than patients suffering from symptomatic nodes (). Comparison of survival times of patients with first relapse classified into ‘doctor-diagnosed’ (n=48) or ‘patient-diagnosed’ (n=77, including symptomatic ‘no follow-up’ patients) did not show significant survival advantages for any group (). Survival analyses were performed in patients attending follow-up in stage III in the same way (‘asymptomatic’ vs ‘symptomatic’; ‘doctor-’ vs ‘patient-diagnosed’) and did not show any survival advantages (data not shown).
Figure 3 Kaplan–Meier curves after detection of symptomatic (symp) and asymptomatic (asymp) first recurrences in stage I/II patients (A) Comparison of survival curves between first relapse patients which were grouped by whether the first recurrence was (more ...)