Premium
Cerebrospinal fluid lactate dehydrogenase isoenzyme analysis for the diagnosis of central nervous system involvement in hematooncologic patients
Author(s) -
Lossos Izidore S.,
Breuer Raphael,
Intrator Orna,
Lossos Alexander
Publication year - 2000
Publication title -
cancer
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 3.052
H-Index - 304
eISSN - 1097-0142
pISSN - 0008-543X
DOI - 10.1002/(sici)1097-0142(20000401)88:7<1599::aid-cncr14>3.0.co;2-j
Subject(s) - medicine , cerebrospinal fluid , lactate dehydrogenase , cart , decision tree , pathology , disease , machine learning , mechanical engineering , biochemistry , chemistry , computer science , engineering , enzyme
BACKGROUND Central nervous system (CNS) involvement is common in hematooncologic diseases. The aim of the current study was to determine the diagnostic value of cerebrospinal fluid (CSF) lactate dehydrogenase (LDH) isoenzyme analysis for the diagnosis of CNS involvement in hematooncologic patients. METHODS The study was comprised of 63 consecutive hematooncologic patients without previous CNS disease who underwent CSF examination as an integral part of their initial staging procedures (44 patients) or for the evaluation of neurologic symptoms (19 patients). Fifteen of these patients had CNS involvement by leukemia or lymphoma. The LDH isoenzyme pattern was established in the CSF of all patients and analyzed by the classification and regression trees (CART) method to construct a decision tree for the prediction of CNS involvement. An additional group of 30 consecutive patients comprised a validation set that was used for cross‐validation of the CART‐derived decision tree. RESULTS A decision tree, with a single split at LDH5 ≥ 2.8% for the prediction of CNS involvement, was constructed and validated by data from a validation set of patients. The decision tree had a sensitivity of 93% and a negative predictive value of 98%. One patient (1.6%) and 2 patients (6.6%) were misclassified in the derivation and validation sets, respectively. Overall, in the combined derivation and validation patient population, the decision tree misclassified 3.2% of patients, whereas CSF cytologic examination misclassified 4.3% of patients. CONCLUSIONS Analysis of the LDH isoenzyme pattern in CSF fluid may be helpful in the evaluation of CNS involvement in patients with hematologic malignancies. The combination of CSF cytology and LDH isoenzyme analysis may improve the sensitivity of CSF cytology significantly. Cancer 2000;88:1599–604. © 2000 American Cancer Society.