Incidence of melanoma recurrence during follow-up

Of all patients with stage I to III completely resected melanomas, 30% develop a recurrence, and approximately 80% of these occur within 3 years of diagnosis of their primary melanoma. Although it is rare for a melanoma to recur after 10 years of follow-up, there are case reports of recurrent melanoma up to 46 years after treatment of a primary melanoma. Late recurrence of melanoma (>10 years after treatment) is largely an observation limited to patients who initially present with early, node-negative primary tumors.

The probability of recurrence is clearly stage dependent. In an analysis of 4748 patients with stage I and II melanoma, melanoma recurred in 18.9% of patients: 5.2% for stage IA, 18.4% for stage IB, 28.7% for stage IIA, 40.6% for stage IIB, and 44.3% for stage IIC.

Patterns of melanoma recurrence during follow-up

In their comprehensive literature review in 2005, Francken and colleagues identified 8 studies that reported the incidence of recurrence in patients with melanoma in follow-up programs. In all patients, most of whom were stage I or II, the rates of recurrence were 3% to 5% for local or in-transit, 5% to 13% for regional nodal, and 3% to 10% for distant recurrence. Among patients whose melanoma recurred, 20% to 28% first present with local or in-transit recurrences, 26% to 60% with regional nodal recurrences, and 15% to 50% with distant recurrences. The wide variation in recurrence rates and patterns can likely be explained by the different proportion of stage I, II, and III patients, as well as how they were staged in these studies. For example, Baughan and colleagues included only stage I patients in their analysis, whereas Hoffman and colleagues included all completely resected stage I, II, and III patients in their series. Most recently, Francken and colleagues from the Sydney Melanoma Unit showed that the site of first recurrence in 211 patients with stage I to III melanoma was local in 13% of patients, in-transit in 17%, regional in 46%, and distant in 24% of patients.

Patterns of melanoma recurrence during follow-up

In their comprehensive literature review in 2005, Francken and colleagues identified 8 studies that reported the incidence of recurrence in patients with melanoma in follow-up programs. In all patients, most of whom were stage I or II, the rates of recurrence were 3% to 5% for local or in-transit, 5% to 13% for regional nodal, and 3% to 10% for distant recurrence. Among patients whose melanoma recurred, 20% to 28% first present with local or in-transit recurrences, 26% to 60% with regional nodal recurrences, and 15% to 50% with distant recurrences. The wide variation in recurrence rates and patterns can likely be explained by the different proportion of stage I, II, and III patients, as well as how they were staged in these studies. For example, Baughan and colleagues included only stage I patients in their analysis, whereas Hoffman and colleagues included all completely resected stage I, II, and III patients in their series. Most recently, Francken and colleagues from the Sydney Melanoma Unit showed that the site of first recurrence in 211 patients with stage I to III melanoma was local in 13% of patients, in-transit in 17%, regional in 46%, and distant in 24% of patients.

Detection of melanoma recurrence

In general, melanoma recurrences are detected more often by patients than by health care professionals at scheduled follow-up visits. Numerous reports in the literature indicate that three-quarters of all recurrences are detected by patients, whereas one-quarter are detected by clinicians. However, some series have found the opposite, with increased rates of recurrences detected by health care professionals. Overall, the rate of first recurrences detected by patients from all reports in the literature is 33% to 99%. Even in stage III patients, who could be considered at highest risk for systemic recurrence and might benefit most from physician-directed follow-up examination and studies, 62% of local and in-transit recurrences, 49% of nodal recurrences, and 37% of systemic recurrences were initially detected by the patients. The retrospective nature of these reports and varied methods of detection of recurrence make comparisons difficult. For example, some studies used questionnaires that were mailed out to all patients whose melanoma recurred and asked them to recall who detected the recurrence, whereas other studies audited the medical record to determine who detected the first recurrence. Yet others based definition of a patient-detected recurrence as one found on an unscheduled clinic visit. In the single prospective study evaluating follow-up of patients with melanoma, only 17% of recurrences were detected by the patients. However, this study captured all recurrences (first and subsequent), so it is difficult to compare with the other retrospective analyses.

A logical question is whether method of detection of recurrence has any effect on survival: no study has ever randomized patients to follow-up versus no follow-up, so level I evidence to address this question is lacking and only indirect observations can be made. Two studies have specifically addressed this question. Mooney and colleagues did not find any difference in overall survival between patients with asymptomatic (recurrences detected by clinicians on physical examination and/or laboratory/imaging investigations) versus symptomatic (patient-detected) recurrences. In their study, 72% of recurrences were detected by patients. Similarly, Francken and colleagues found no difference in survival in recurrences detected by patients versus recurrences detected by physicians. In their report, 73% of recurrences were detected by the patients, and most were symptomatic. In contrast, Poo-Hwu and colleagues reported a 5.8% improvement in overall survival for patients with asymptomatic recurrences discovered at routine follow-up, compared with symptomatic recurrences discovered by patients. However, biases in lead time (earlier detection that increases the time between diagnosis and death artificially) and length time (slow-progressing lesions with less aggressive potential are preferably detected), differences in the biologic nature of tumors discovered by patients versus health care providers, and differences in patients included in these analyses could account for these differences. In their prospective evaluation of melanoma recurrence, Garbe and colleagues stratified first recurrences as either early (asymptomatic) or late (symptomatic) and found a higher survival rate for patients with early versus late recurrences (76% vs 38% overall survival after 3 years). In this analysis, recurrences were not stratified based on who detected them (patient vs health care professional), and patients with all types of recurrences (first or subsequent) were included in their study and evaluation, once again limiting the ability to compare these results with other studies. Romano and colleagues specifically addressed site and timing for first relapse in stage III patients with melanoma. Their evaluation of 340 patients with a median follow-up of 77 months found overall recurrence-free survival rates for stage IIIA, IIIB, and IIIC of 63%, 32%, and 11%, respectively. Site of first relapse was local/in-transit in 28% of patients, regional nodal in 21% of patients, and systemic in 51% of patients. First relapses were detected by the patient in 47% of cases, clinically by the physician in 21% of patients, and by screening radiologic tests in 32% of patients. Multivariate analysis found that improved survival was associated with younger age, local/in-transit or nodal recurrence, asymptomatic recurrences, or resectable systemic recurrences. The investigators concluded that, based on the timing and site of relapse, routine physical examinations beyond 3 years for stage IIIA, 2 years for stage IIIB, and 1 year for stage IIIC and routine radiologic imaging beyond 3 years for stages IIIA and IIIB and 2 years for stage IIIC patients was of low yield and that published guidelines should strongly consider whether or not they should be included in recommendations for long-term follow-up.

Frequency of follow-up

Guidelines from most countries, melanoma cooperative groups, and consensus guidelines recommend follow-up visits with a physician who specializes in melanoma (surgical oncologist, dermatologist, or medical oncologist) with a complete history, physical examination (including skin and regional lymph node examination) every 3 to 12 months for the first 5 years after treatment of melanoma. Generally, follow-up is spaced out to every 6 to 12 months for 5 more years, with stage-specific frequency of visits. Table 1 summarizes the recommendations from 3 major multidisciplinary groups with expertise in melanoma (the National Comprehensive Cancer Network [NCCN, United States], the Clinical Practice Guidelines for the Management of Melanoma in Australia and New Zealand, and the German Cancer Society) across 3 continents. The variability in follow-up frequency in these 3 recommendations is representative of other national and professional organizations.

Based on the consistent observations that most recurrences are detected by patients and that patient- versus clinician-detected recurrences do not affect survival, the concluding remarks from many studies suggest less frequent follow-up, although a specific time course outside major published guidelines is rarely suggested. For example, Francken and colleagues conclude that “the frequency of follow-up visits in an earlier SMU (Sydney Melanoma Unit) study and by other groups could be reduced for most patients.” They go on to remark, “However, individual patient-related factors such as physical and mental health, travel distance, and personal preferences will always need to be considered, and the extent of resources available to conduct follow-up programs may also be of relevance.” This statement summarizes the sentiment that many melanoma providers instinctually feel: that increased frequency follow-up does not affect survival and in the era of burgeoning practices and cost constraints, it is difficult to justify. A follow-up article recommends annual follow-up for stage I patients, 6-month follow-up for 2 years, followed by annual visits for stage IIA patients, and 4-month follow-up for 2 years, 6-month follow-up in year 3, and annual follow-up thereafter for stage IIB and IIC patients. As discussed in the preceding section, follow-up examinations and imaging tests were found to be of low yield beyond 3 years in patients with resected stage III melanoma.

In their prospective evaluation of follow-up, Garbe and colleagues adhered to the guidelines published by the German Society of Dermatology, which recommended follow-up examinations every 3 months in the first 5 years, with continued follow-up every 6 months until the 10th year. Based on the results of their study, the German Cancer Society adopted the guidelines presented in Table 1 , which slightly increased the follow-up intervals for early-stage melanoma.

One strategy to continue high-frequency follow-up, but relieve the burden of the melanoma specialist provider, is to transition patients to their primary care physician (general practitioner [GP]) for follow-up. A randomized trial of GP-led melanoma follow-up in the United Kingdom showed that such a strategy is safe, feasible, engenders greater patient satisfaction, permits closer adherence to guidelines, and does not result in adverse effects on health status or anxiety. There were 142 patients enrolled in this study over a 1-year period. Participating GPs were educated with a 4-hour training session and comprehensive literature for follow-up care, which focused on presentation of new and recurrent melanomas. Although the numbers are small, this study indicates that melanoma follow-up can be safely provided by any clinician, including physician extenders (such as nurse practitioners or physician assistants) educated in the timing and patterns of melanoma recurrence. This finding should not be surprising given the high likelihood that most recurrences are detected by patients.

Usefulness of laboratory and radiographic tests in the follow-up of the patient with melanoma

In the preceding sections, it seems that patient versus clinician detection of melanoma recurrences, along with the timing and frequency of follow-up visits, does not have an effect on survival. A logical question, then, is whether the addition of laboratory or imaging tests to clinic visits may be able to detect subclinical recurrences, which could then be treated with a combination of surgery, chemotherapy, and/or radiation therapy to improve outcomes. A central premise to this question is that earlier detection of recurrence improves survival. Although intuitively, oncologists often believe this hypothesis to be the case, there is no high-level evidence in melanoma to support it. Intensive follow-up with frequent laboratory and imaging studies, which have a significant false-positive rate, may be detrimental in terms of cost and patient-related anxiety. The next section reviews the evidence for inclusion of such studies in melanoma follow-up programs.

Radiographic Tests

The role of routine radiographic testing as an adjunct to scheduled clinical visits to detect early recurrences follows the aforementioned unproven hypothesis that earlier detection of melanoma recurrences leads to improved survival. Although effective treatment of local, in-transit, and regional node metastases offers the possibility of cure, long-term survival after resection of systemic metastases is less frequent, but is possible. There are reports using chest radiograph (CXR), regional/nodal ultrasound (US), abdominal US, and computed tomography (CT) and magnetic resonance imaging (MRI) scans in the follow-up of patients with melanoma. Similar to studies mentioned earlier, the comparison of these studies is difficult given the heterogenicity of the patient cohorts, time frames of analysis, and definitions of recurrences. Further, with the exception of the single prospective study by Garbe and colleagues, all of these studies are retrospective in nature. However, the following studies represent the best data to date to guide our clinical practice. Each imaging modality is discussed in the following sections.

CXR and chest CT to detect pulmonary metastases

Screening for pulmonary metastases with CXR and CT scans in patients with melanoma is based on the observation that the lungs are the most common visceral site of metastases. Of patients with pulmonary involvement, 12% to 25% are eligible for surgical resection, with 80% to 90% of these patients undergoing complete (R0) resection. The 5-year survival in this highly selected group can be as high as 30%. Thus, the goal of radiographic screening tests to detect pulmonary metastases is to detect asymptomatic lesions at an earlier stage that are amenable to an R0 resection. Ideally, a randomized trial of routine pulmonary imaging versus no imaging could answer the question of whether these tests improve survival; however, these studies do not exist. CXR is the most studied test to date. In terms of detection of pulmonary metastases, there are mixed findings in the published literature concluding that CXR is able or unable to detect asymptomatic pulmonary metastases from melanoma. In the studies that found that CXR was able to detect pulmonary recurrences, the rate was 0.5% to 6.2% of patients. The wide variability reported in these studies can be explained by the wide range of patients included and the differing length of follow-up. For example Kittler and colleagues and Bassères and colleagues specifically analyzed patients with thin melanomas, whereas studies by Kersey and colleagues, Goerz and colleagues, and Mooney and colleagues each included more high-risk stage II and III patients in their analyses. No study has shown that the early detection of pulmonary metastases by routine imaging is associated with improved survival. Taso and colleagues examined 994 patients at the Massachusetts General Hospital and found that 75 (8%) patients had pulmonary metastases detected by CXR (1937 total CXRs). In addition, there were 63 (6% of patients) with false-positive results. In 41 patients (4%), CXR was the initial evidence of metastases. Survival after identification of pulmonary metastases did not differ between patients with asymptomatically detected pulmonary metastases versus those with symptomatic or pulmonary metastases found in the setting of other metastatic sites. Meyers and colleagues found that in 118 patients with stage II or III melanoma followed with routine clinical examinations and imaging tests over at least 2 years, only 2 of 43 recurrences (5%) were detected by CXR. At the SMU, Morton and colleagues enrolled 108 patients with stage III patients in a prospective monitoring schedule of CXR every 6 months for 5 years. Patients were followed for a median of 52.5 months, and 23 (21%) developed pulmonary metastases, of which 11 (48% of recurrences) were detected by CXR. Three of the 23 patients with pulmonary recurrences detected by CXR (13% of total) went on to resection. Further, abnormal CXRs were found in 19 additional patients, but not because of recurrence (false-positive result). Median survival did not differ between recurrences detected by CXR and those not detected by CXR. The investigators concluded that because routine CXRs detected only half of pulmonary recurrences, infrequently identified patients for potentially curative surgery, had a high rate of false positivity, and did not affect survival, their use was not justified. In Garbe and colleagues prospective study of melanoma follow-up, CXR detected metastases in 2 patients (11.1%) with stage I melanoma, 11 patients (22.4%) with stage II melanoma, and 13 patients (9.5%) with stage III melanoma. Overall, 0.6% of all CXRs were responsible for the detection of metastases, and as many as 75% of recurrences detected by CXR were first discovered at an advanced phase. These investigators concluded that the routine CXR was of little benefit for stage I and II patients and should be restricted to follow-up of patients with more advanced stages of disease.

Mooney and colleagues performed a cost-effectiveness analysis using data from the Roswell Park Cancer Institute and the National Cancer Institute’s Surveillance, Epidemiology, and End Results (SEER) program. They found that the cost of CXR screening per nonquality-adjusted life year and quality-adjusted life year were $150,000 and $165,000, respectively, in 1996 US dollars. This amount accounted for 80% of program costs. Based on these findings, the investigators suggested reducing the frequency of screening CXR.

In an effort to improve the usefulness of CXR in detecting pulmonary recurrences of melanoma, 3 studies have evaluated the use of chest CT scans. Only one identified metastases, but in only 3 of 364 patients with stage I to III melanoma, 2 of whom had distant metastases elsewhere. However, these studies are limited; the study by Bassères and colleagues evaluated only patients with stage I melanoma, whereas the study by Mijnhout and colleagues comprised only 67 patients with resected stage I to III melanoma, 51 of whom (76%) underwent screening CT scans. Given these results, along with the inability of screening CT scans to readily detect primary lung cancers in heavy smokers without an unacceptable rate of false-positive results, this modality does not seem appropriate for general use in the follow-up of the patient with melanoma.

The overall sensitivity, specificity, false-positive, and false-negative rates of studies examining screening CXR in the follow-up of patients with melanoma is presented in Table 2 . The inability of this testing modality to detect recurrences and affect survival has led to the omission of routine CXR in the NCCN and Australia/New Zealand guidelines and the limited use of CXR only in stage II and III patients in the German guidelines (see Table 1 ).

Table 2

Only gold members can continue reading. Log In or Register to continue

Share this:

• Click to share on Twitter (Opens in new window)
• Click to share on Facebook (Opens in new window)

Related

Related posts:

Melanoma Pathology: Important Issues for Clinicians Involved in the Multidisciplinary Care of Melanoma Patients Surgical Approach to Primary Cutaneous Melanoma Sentinel Lymph Node Biopsy for Melanoma: Critical Assessment at its Twentieth Anniversary Immunotherapy of Melanoma: An Update Targeted Therapy for Melanoma: A Primer Assessment of Patient-Reported Outcomes in Patients with Melanoma

Stay updated, free articles. Join our Telegram channel

Join