Detection of occult metastatic melanoma by urine chromatography.

By using ion-exchange column chromatography with effluent monitoring using the stable, free radical alpha,alpha-diphenyl-beta-picryhydrazyl as a colorimetric reagent, we have demonstrated the occurrence of elevated levels of five peaks in the urine of patients with metastatic disease. The tentative assignment of two of the peaks as 3,4-dihydroxyphenylalanine and as 3-methoxy-4-hydroxyphenylalanine has been made. Three remain unknown. The correlation of these peaks with the clinical status of melanoma patients shows that, while the individual excretion pattern of these compounds may be variable, the sustained occurrence of one or more of them in a patient's urine is evidence of recurrent or continuing disease. The excretion levels appear to be proportional to the tumor burden. The results with a group of 39 melanoma patientshaving Stage II or Stage III disease indicate that this chromatography technique provides earlier evidenc eof liver metastases than doses the liver scan, may detect occult metastases generally, and has detected tumor in clinically enlarged lymph nodes. This method, in its present form, does not detect small pulmonary lesions earlier than chest X-ray or tomography do or brain metastases earlier than do brain scan or computerized axial tomography. The technique is clinically useful for the diagnosis of melanoma patients and in their follow-up while under treatment.

Lentigo maligna melanoma has no better prognosis than other types of melanoma.

We studied 48 patients with lentigo maligna melanoma (LMM) and compared the clinical stage I patients with non-LMM melanoma patients (matched by site and thickness) to see if prognosis differed. There was no significant difference in mortality from melanoma between the two groups (P = .68) after a mean follow-up time of five years (67.5 months for LMM, 60.5 months for non-LMM). In addition, a Cox multivariate analysis of the entire matched group showed that only thickness was significantly associated with death from melanoma (P = .0007) while histology (LMM v non-LMM) did not make a significant contribution (P = .61). Our data suggest that after accounting for primary tumor thickness and site, LMM and non-LMM have the same prognosis and biologic behavior, in contrast to the widely held belief that LMM has a better prognosis than other forms of melanoma.

The relationship of psychosocial factors to prognostic indicators in cutaneous malignant melanoma.

This study investigated the relationship between prognosis (estimated by histopathologic indicators) in cutaneous malignant melanoma and a comprehensive set of physical risk, demographic, psychosocial, and situational variables. These variables were derived from the medical examination, the pathology report, psychosocial self-report measures, and an hour-long videotaped interview with 59 patients from two melanoma clinics in San Francisco. Variables significantly correlated with tumor thickness were: darker skin/hair/eye coloring, longer patient delay in seeking medical attention, two correlated dimensions within an operationally defined 'Type C' constellation of characteristics, two character style measures, and less previous knowledge of melanoma and understanding of its treatment. Of these variables, delay was the most significant in a hierarchical multiple regression analysis in which tumor thickness was the dependent variable. Associations between tumor thickness and psychosocial measures of Type C were considerably stronger and more significant for subjects less than age 55, suggesting that the role of behavioral and psychosocial factors in the course of malignant melanoma is more potent for younger than for older subjects.

Melanoma: value of early detection and treatment.

Recent studies confirm that the prognosis in patients with melanoma depends on the level of invasion by the tumor at diagnosis. Superficial spreading melanomas usually are detected during the initial radial growth phase, when simple excision usually effects a cure. Nodular melanomas extend vertically from inception and often are not detected until they have penetrated to relatively deep levels. Because the possibility of recurrence is higher with deeper tumors, lymph node dissection may also be necessary. Fortunately, superficial spreading melanomas are considerably more common than the nodular type. Therefore, increased efforts at earlier detection and diagnosis of melanoma will be rewarded by increased cure rates.

Clinical judgment and computers.

The increasingly frequent application of formal methods, including algorithms and computer programs, to processes that are ordinarily viewed as judgmental seems to be a source of both promise and unease for physicians. A consideration of some of these methods suggests that it may be helpful to attempt to distinguish carefully between judgment and computation. Medical care involves a complex of inferential processes, any of which may be performed as judgments, and some of which may be carried out as a computation. My purpose here is to identify the latter cases. The empirical evidence suggests that such a demarcation is feasible. The most important question appears not to be "Where can we use computers?" but "Where must we use human beings?" Until this matter is more thoroughly explored, tension between physicians and computer advocates will persist.

Conceptual issues in computer-aided diagnosis and the hierarchical nature of medical knowledge.

Attempts to formalize the diagnostic process are by no means a recent undertaking; what is new is the availability of an engine to process these formalizations. The digital computer has therefore been increasingly turned to in the expectation of developing systems which will assist or replace the physician in diagnosis. Such efforts involve a number of assumptions regarding the nature of the diagnostic process: e.g. where it begins, and where it ends. 'Diagnosis' appears to include a number of quite different cognitive processes, some of which seem more subject to formalization than others. Underlying the difficulties inherent in these efforts at formalization, is the hierarchical structure of medical knowledge including that of diseases. (By 'hierarchical structure', I mean no more than that our descriptions of clinical objects necessarily refer to different levels of organization; atomic, molecular, cellular, physiological system, entire patient). Since diagnosis is in part a form of classification, measures of similarity between the clinical findings of a particular patient and a potential disease, are needed. This is complicated by the fact that some diseases are described using the vocabularies of physics and chemistry, others with the vocabulary of psychology, and most, with terms taken from a number of different hierarchical levels. The differences between the descriptions made using these vocabularies will be seen to exert systematic (if not uniform) influences upon the diagnostic process.

Physical studies on melanins. II. X-ray diffraction.

A number of purified natural and synthetic melanins have been examined by X-ray diffraction. A consistent finding with all samples was the lack of structure in the diffraction pattern corresponding to any significant crystallinity in these melanin preparations. A diffuse ring, centered at a Bragg spacing of 3.4 A was consistently found in samples of melanin from animal sources, and a similar ring at 4.2 A in all melanins obtained from plants. Models for these two polymer types, based upon the current concept that they primarily involve indole and catechol monomeric units respectively, were then evaluated by a Monte Carlo method. From the comparison of the observed spacings with the calculated ones it was concluded that the 4.2 A spacing in the catechol melanins is probably related to the average interaction between adjacent monomeric units, with mutually random orientations. The 3.4 A spacing observed in indole melanins appears to derive from the tendency of indole monomers (probably of adjacent chains) tending to aggregate in near parallel stacks. Some randomness in the form of translations and rotations parallel to the planar groups is consistent with the diffraction patterns. An interesting finding was that the diffraction pattern of synthetic melanin prepared by the alkaline auto-oxidation of catechol gave the 3.4 A spacing found in the indole melanins of natural origin.

The regressing thin malignant melanoma: a distinctive lesion with metastatic potential.

To validate the supposition that thin malignant melanomas (less than 0.76 mm thick) of ordinarily low risk but with areas of regression may paradoxically metastasize, we observed 121 thin malignant melanomas over a six year period. Of these, 23 displayed readily apparent areas of regression, of which five (21.7%) metastasized. The incidence of metastases in their 98 counterparts without regression was 2.0% (2/98). The difference between the two is statistically significant (p = less than .01). Of the entire group of the two is statistically significant (p = less than .01). Of the entire group of thin melanomas, those with regression represented 19.0% (23/121) yet accounted for a disproportionate 71.4% (5/7) of all metastases. We conclude that regression is a relatively poor prognostic sign, whose occurrence within an otherwise thin melanoma represents a significant caveat to the current histologic staging system that equates thinness with low risk. We thus submit that patients whose malignant melanomas display regression be followed rigorously for evidence of metastases irrespective of the tumor's actual measured thickness or level of invasion.

Regression in malignant melanoma. A histologic feature without independent prognostic significance.

A total of 844 cutaneous malignant melanomas were examined prospectively for the presence or absence of histologic regression within the primary tumor. Cases were then stratified into three groups according to tumor thickness and survival was compared between substrata with and without regression in each group. The distribution of other major prognostic variables within these substrata was assessed and their influence as potential confounding variables considered. No statistically significant effect of regression on survival was found in any of the three thickness strata. These results do not confirm the finding of an earlier study, which suggested that regression may be a poor prognostic sign when found in association with thin malignant melanomas. Regression was almost invariably associated with the radial growth phase of melanomas. Regression was more common in male than in female patients, and was more frequent in association with truncal than extremity or head and neck melanomas.

Frequency and duration of patient follow-up after treatment of a primary malignant melanoma.

To develop guidelines for the follow-up of patients with primary cutaneous melanoma (clinical Stage I), we studied 295 patients who had presented with a primary melanoma and who subsequently developed evidence of metastatic disease in the course of follow-up. Cox multivariate analysis was used to assess the influence of five variables in predicting the interval of time from the diagnosis of melanoma to the first clinical or laboratory evidence of metastatic disease (disease-free interval). The variables studied were tumor thickness, patient sex, patient age, elective lymph node dissection, and primary tumor location. Tumor thickness was found to be the major predictor of disease-free interval, which shortened progressively with increasing tumor thickness. Men had a shorter mean disease-free interval than women, although this effect did not reach statistical significance at the 0.05 level. Patient age, tumor location, and elective lymph node dissection were found not to be predictors of disease-free interval. The risk of recurrence of melanoma was tabulated, by year, for four intervals of tumor thickness. The increase in risk of recurrence associated with increases in tumor thickness above 1.5 mm was shown to occur predominantly in the early years following diagnosis-particularly in the first year. On the basis of our findings, we have suggested regimens of follow-up for melanoma.

Column cation-exchange separation of melanin-related metabolites in urine from cases of melanoma.

We describe recent developments in the use of a stable free radical, diphenylpicrylhydrazyl, as a colorimetric detector of reducing metabolites in melanoma urines. A procedure for buffering aqueous/solvent reagent solutions is presented, and examples of the baseline stability achieved for reference chromatograms and urine samples are provided. Chromatograms of phaeomelanin precursors and of an extract of a highly pigmented hamster melanoma are also presented. Identities are tentatively assigned for some of the chromatographic peaks that have previously been correlated with disease, including isomers of cysteinyldopa, and observations of new pigment- and tumor-related metabolites in the chromatograms are noted.

Neural network approach to detection of metastatic melanoma from chromatographic analysis of urine.

Chromatographic analysis of sera or urine is important in medicine for the evaluation of patients whose clinical status is associated with the presence of specific biochemical markers. Malignant melanoma has been a model for such studies due to the elaboration of melanin precursors and pigment as the tumor metastasizes. Computer-assisted methods for categorizing chromatographic data and clinical status are imperative due to the large number of detectable compounds and possible correlations. In addition, computer-based analysis of the data can readily extract patterns that are not obvious by visual inspection. In this paper, we present a neural network analysis of melanoma chromatographic and clinical data that categorizes subjects into normals, NED patients (No Evidence of Disease), and metastatic patients. The set of marker compounds for metastatic disease represents a significant advance over the correlations derived by visual inspection.