Diagnostic tests and Hodgkin's disease.A standardized approach to their evaluation

Abstract

We have analyzed a series of 124 young men with Hodgkin's Disease in an attempt to clarify the role of non-invasive tests in the staging evaluation. Data are presented in a way which permits comparison between series and application of the data to decision-making models. Analysis of the data shows our patients to have had primarily limited (stage I and 11) disease. In our hands certain tests were quite accurate in predicting the absence of abdominal disease (lymphangiography, hepatic scintigraphy) as determined by laparotomy, but no study appeared to eliminate the need for laparotomy in this group. EVERAL RECENT REPORTS HAVE EMPHASIZEDS the role of decision analysis in the initial management of patients with Hodgkin's Disease. 16917 This type of approach, while theoretically attractive, suffers because data have not been presented in the medical literature in a way that permits their incorporation into a decision analysis model. There have been, for instance, no studies of the expensive, non-invasive diagnostic tests usually performed on Hodgkin's patients where the data were analyzed by methods usually recommended to evaluate a diagnostic test. Such methods include presentation of the sensitivity and specificity of the test along with the "prior probability" of disease in the population under study. of diagnostic tests, along with data on the population under study are available, then the efficacy of any diagnostic test can be rationally evaluated using standard techniques. Such data also permit more intensive scrutiny of inter-institution variability and facilitate testing of decision-making models.We have evaluated the characteristics of several non-invasive tests in a large, untreated population of patients with Hodgkin's disease. By comparing test results with operative findings we have determined the sensitivity and specificity of these tests vis a' vis abdominal disease as defined by laparotomy. Additionally, our data are based solely on young men. This permits determination of the "prior probability" of abdominal involvement in this group. Data from our series can now be compared with data from different patient populations.We have also examined the other large series in the literature which report data on non-invasive tests in Hodgkin's disease. T o encourage other authors to present data in this form we have determined, when possible, test sensitivities and specificities. We have also presented data on the "predictive accuracy" of the diagnostic tests. The clinically important "predictive accuracy'' can be calculated for any test on any given patient if the sensitivity, specificity, and prior probability are known. These data can now be combined with other clinical parameters and their effect on treatment and survival measured.

Assessment of biochemical tests for bone and liver involvement in malignant lymphoma patients

“…For the detection of hepatic involvement in the malignant lymphomas, 5NT is more reliable and has been reported by one group to be more accurate for the diagnosis of metastatic liver disease than liver scanning.24 A more recent report, in which radiological data from four centers were summarized, showed that hepatic scintigraphy correlated poorly with results of laparotomy in detecting hepatic involvement in patients with Hodgkin's Disease. 20 Other authors found an unsatisfactory correlation between baseline liver tests, liver biopsy results, and the clinical course of patients with Hodgkin's Disease (predominantly Stage 1-11) followed up for a minimum period of four years, but the biochemical tests employed were restricted to aminotransferases, Alp, bilirubin, and bromosulphalein clearance ,21 moreover, the methods and reference ranges were not described in their report. As shown in Table 5, 14 cases with initial or subsequent hepatic involvement gave abnormal serum 5NT activities, whereas detection rates for the other four enzymes ranged from 7-9.…”

Staging laparotomy with splenectomy for Hodgkin's disease: The Stanford experience

“…Harris et al [19] analyzed 124 lymphangiograms from Walter Reed Hospital and found the predictive value of a positive test to be 36%, and negative tests had a predictive value of 91%. This is equivalent to the terms accuracy positive/negative used in this report.…”