The Prognostic Significance of Pretreatment Serum Lactate Dehydrogenase in Patients With Small-Cell Lung Cancer By U. Sagman, R. Feld, W.K. Evans, D. Warr, F.A. Shepherd, D. Payne, J. Pringle, J. Yeoh, G. DeBoer, A. Malkin, and R. Ginsberg Pretreatment serum lactate dehydrogenase (LDH) levels were assayed in 288 patients presenting with small-cell lung cancer (SCLC) between 1976 and 1985. Patients were routinely staged by physical examination, chest x-ray, bone, brain, and liver scans, and bone marrow evaluation. Clinical response and survival were assessed following treatment with combination chemotherapy as part of four clinical trials. Patients with extensive disease (ED) presented with a higher incidence (108 of 147, 73%) of abnormally elevated LDH (> 193 IU/L) than those (65 of 141, 46%) with limited disease (LD) (P = 2 x 10-6). Forty percent of patients had an initial normal LDH level and a higher response rate (89 of 108, 82%; complete response [CR], 47%) than those with elevated values of LDH (119 of 156, 76%; CR, 29%). The CR rate varied inversely with the level of LDH in patients with LD (P = .026) but not in those with ED (P = .300). The median survival time and 1-year and 2-year survival
rates for patients with elevated LDH were 39 weeks and 33%and 6%, respectively, whereas for those with a normal LDH level these were 53 weeks and 54% and 16%, respectively. Patients with LD and elevated levels of LDH manifested a higher relative death rate (1.63:1)
THE
measure of individual outcome as well as optimize patient stratification into clinical trials and therefore lead to improved assessment of such studies. The most important prognostic factors identified to date have included extent of disease and clinical performance status (PS). 5 The prognostic importance of several biologic markers has also been assessed in patients with SCLC. To date, most substances examined have lacked stringent sensitivity and specificity as diagnostic adjuncts. Pretreatment serum levels of biologic markers, such as the carcinoembryonic antigen (CEA) and neuron-specific enolase (NSE), however, are important determinants of prognosis6 ' 7 and hence may have a role in the design of treatment plans used to treat patients with SCLC. The possible role of the serum enzyme lactate dehydrogenase (LDH) as a prognostic attribute has recently been examined in patients with lymphoma,' 9 selected histologic types of sarcoma,1 0 and multiple myeloma., 1 ' 2 We have measured the level of serum LDH in patients with SCLC before treatment on clinical studies at our institutions. We report our analysis of the value of LDH in the staging and prognosis of patients with SCLC.
SURVIVAL of patients with small-cell carcinoma of the lung (SCLC) has improved with the advent of new chemotherapeutic agents and the use of combination chemotherapy.' Without treatment, the median length of survival after diagnosis is 1 to 3 months.2 Combination chemotherapy produces objective tumor regression in as many as 90% of patients and median survival times of approximately 8 to 11 months for patients with extensive disease (ED) and more than a year for those with limited disease (LD).3 4 With this improved outcome, it is desirable to identify prospectively patients with a more favorable prognosis. This knowledge may provide a more accurate
From The Ontario Cancer Institute Incorporating The Princess Margaret Hospital, Toronto General Hospital, Mt Sinai Hospital, Sunnybrook Medical Center, University of Toronto, Toronto; Ontario Cancer Treatment and Research Foundation; and Ottawa Regional Cancer Centre, Ottawa, Canada. SubmittedApril 9, 1990; accepted December 26, 1990. Address reprintrequests to Ronald Feld, MD, The Ontario Cancer Institute, 500 Sherboume St, Toronto, Ontario M4X 1K9 Canada. 0 1991 by American Society of Clinical Oncology. 0732-183X91/0906-0006$3.00o/0
954
when compared with patients with LD and LDH in the
normal range (P = .0083). The survival of patients with ED did not differ between those with normal and elevated levels of LDH (P = .273). A significant survival advantage persisted for patients with LDH in the normal range following adjustments for extent of disease, performance status (PS), and treatment protocol (P = .044, log-rank analysis). In conclusion, serum LDH appears to be a significant independent pretreatment prognostic factor in patients with SCLC that correlates with stage of disease, response to treatment, and survival. J Clin Oncol 9:954-961. o 1991 by American Society of ClinicalOncology.
Journalof Clinical Oncology, Vol 9, No 6 (June), 1991: pp 954-961
Downloaded from ascopubs.org by University of Newcastle Upon Tyne on April 7, 2019 from 128.240.208.034 Copyright © 2019 American Society of Clinical Oncology. All rights reserved.
955
PROGNOSTIC SIGNIFICANCE OF LDH IN SCLC MATERIALS AND METHODS All patients with SCLC managed at University of Toronto affiliated hospitals from December 1976 to December 1986 were considered for retrospective analysis in this study. All patients had histologic or cytologic proof of SCLC, had received no prior treatment with chemotherapy, radiotherapy, or surgery, and had a performance status (PS) from 0 to 3 on the Eastern Cooperative Oncology Group (ECOG) scale. They were excluded if they had a history of previous malignancy, unless disease-free for at least 5 years, or if the tumor was a nonmelanoma skin cancer. Patients were free of active cardiac disease, severe hepatic dysfunction (serum bilirubin > 2 mgi100 mL), and had normal renal and hematologic function. Those patients included in a study of total parenteral nutrition as an adjunct to chemotherapy required (see below) assessment for evaluation of disease and freedom from CNS metastases, serious uncontrolled infection, superior vena cava obstruction, hyponatremia (< 125 mEq/L), or severe hypoalbuminemia (< 2.5 g/100 mL). Patients on all other studies did not have these exclusions. All patients had a complete history, physical examination, evaluation of performance status, and investigations to define the extent of their disease, which included routine hematology and chemistry, chest x-ray, bone, liver, and spleen radioisotope scans, and bone marrow aspirate and biopsy. Patients with CNS symptoms and most other patients had pretreatment brain scintiscans or computed tomographic scans. Those patients with clinical evidence of tumor confined to a hemithorax and ipsilateral supraclavicular nodes (inclusive of cytology-negative pleural effusions) were considered to have LD. All others with evidence of disease spread beyond the extent defined for LD were designated to have ED. Pretreatment serum LDH determinations were performed on 288 of 614 patients who were managed during the 10-year study interval. Serum LDH determinations were not part of the routine pretreatment investigations, but were performed randomly at the discretion of attending physicians as part of a battery of tests aimed at evaluating hepatic dysfunction specifically and overall disease activity in general. No other tests were performed as a direct result of elevated LDH levels during the pretreatment interval, and therefore, the staging investigations of the study population were not modified. The proportion of patients on whom LDH levels were determined as a function of year of study was relatively uniform throughout the period covered by the study. The two hundred eighty-eight patients on whom serum LDH levels were evaluated form the basis of this report. Two hundred forty of these patients were managed in controlled clinical trials at our institutions along with other patients on whom LDH levels were not determined. The remaining 48 patients were not included in a specific trial but had identical investigations and staging procedures. One hundred fifty-one patients were derived from previously reported study" in which initial treatment consisted of 2 cyclophosphamide 900 mg/m , doxorubicin (Adriamycin; Adria Laboratories, Columbus, OH) 45 mg/m2, and vincristine 2 mg by bolus intravenous (IV) injection (CAV).
Treatments were repeated at 3-week intervals for a total of six courses unless tumor progression was evident. Approximately 1 week following the third course of chemotherapy, the responding patient underwent whole-brain prophylactic irradiation (2,000 cGy in five equal fractions). Following six cycles of CAV, the pretreatment volume of intrathoracic tumor was irradiated in responding patients with both LD and ED (2,500 cGy in 10 equal fractions over 2 weeks). At relapse, re-treatment with the CAV regimen was generally attempted, and if reinduction failed, further treatment was administered at the discretion of the attending physician. 4 The 43 patients, derived from a second published trial,' were treated in a randomized fashion to evaluate the efficacy of total parenteral nutrition (TPN) as an adjunct to chemotherapy (part of a multicenter National Cancer Institute-sponsored trial). In this study, patients received two courses of CAV, which consisted of cyclophosphamide 2 1,200 mg/m 2, Adriamycin 70 mg/m , and vincristine 2 mg administered IV by bolus injection at 3-week intervals. Those achieving an objective response received two further courses of lower dose CAV followed by prophylactic cranial irradiation (2,000 cGy in five equal fractions), and in those with LD, chest irradiation was administered (2,500 cGy in 10 equal fractions over a period of 2 weeks). Patients were then treated with etoposide (VP-16) 100 mg/m2 for 3 days and methotrexate 50 mg/m2 IV on day 1 with leucovorin (VMC) every 3 weeks for five cycles. Treatment with further 2 cycles of CAV to a total Adriamycin dose of 450 mg/m was followed by five additional courses of VMC. Patients who did not respond to the two initial courses of CAV were treated with VMC to disease progression. Patients randomized to receive TPN in the second protocol were started on nutritional support 7 days before the initial course of chemotherapy, and TPN was discontinued 48 hours following the administration of the second cycle of CAV. A third group of patients were obtained from two trials, (combined here as BR), which evaluated the efficacy of two alternating noncross-resistant chemotherapeutic regimens 5 in the management of patients with ED, BR ,'4 and LD, 6 BR,.' In the BR 4 trial, 46 patients were randomized to receive either six courses of IV cyclophosphamide 1,000 2 mg/m 2, doxorubicin 50 mg/m , and vincristine 2 mg (CAV) at 3-week intervals, or CAV alternating with VP-16 100 mg/m2 2 days 1 to 3 and cisplatin 25 mg/m on days 1 to 3 every 3 who progressed on management of patients weeks. Further the alternating regimen was instituted at the discretion of the physician. The four patients derived from the BR 3 trial were randomized to receive three courses of CAV followed by three courses of VP-16 and cisplatin for a total of six courses. All responders received prophylactic cranial irradiation (2,000 cGy in five equal fractions) following three cycles of treatment and were randomized for locoregional control with either standard (2,500 cGy) or high-dose (3,750 cGy) irradiation regimens. Forty-eight patients were not on study (NOS) protocols, and comprised those not meeting eligibility criteria for concurrently active trials. The characteristics of the 48 patients including age, sex distribution, performance status, and disease extent were similar to those of the patients who did enter study protocols. Moreover, as a group, their response to treatment and overall survival was not signifi-
Downloaded from ascopubs.org by University of Newcastle Upon Tyne on April 7, 2019 from 128.240.208.034 Copyright © 2019 American Society of Clinical Oncology. All rights reserved.
956
SAGMAN ET AL
cantly different from those who were managed according to clinical trial protocols. Although management of such patients was based on the investigator's option, most received a chemotherapeutic regimen that included cyclophosphamide, Adriamycin, and vincristine (a modified CAV protocol). Criteria for response to treatment were defined uniformly in all patients in this study. Complete response (CR) was defined as disappearance of all known neoplastic disease. A partial response (PR) was defined as a decrease > 50 in the product of the two greatest perpendicular diameters of well-outlined lesions, lasting for at least 3 weeks in the absence of progressive disease elsewhere or the development of any new lesions. In the assessment of poorly defined lesions, an estimated 75% decrease in the cross-sectional area was required to qualify for a PR, again in the absence of progressive disease elsewhere or the development of any new lesions. Progression of disease was defined as a greater than 25% increase in the cross-sectional area of one or more lesions or the occurrence of any new lesions. Survival time was measured from the date of the first chemotherapy administration to the date of death. LDH Determinationand Normal Range Pretreatment serum LDH activity was assayed by the spectrophotometric absorbency of nicotinamide-adenine dinucleotide at 340 nm, following catalytic oxidation of lactate to pyruvate. The normal reference range based on 276 apparently healthy subjects was considered to be between 100 to 193 IU/L (ACA, Du Pont automatic clinical analyzer). No isoenzyme fractionation was undertaken. LDH assay determinations were performed in the respective departments of Clinical Biochemistry at The Princess Margaret Hospital and the Toronto General Hospital (Toronto). StatisticalConsiderations Survival curves were calculated by the method of Kaplan and Meier.17 The log-rank method"' was used to test the statistical significance of the difference between survival curves, as well as to do comparisons adjusted for other prognostic factors. P values for differences between proportions were calculated with Fisher's exact test (two-tailed). Correlations between LDH level (normal or elevated) and specific metastatic sites were assessed with the product moment correlation coefficient (r). RESULTS Of the total 614 patients managed during the study interval, those without (326 patients) and
overall response and CR rates for those with LDH determination were 79% and 36%, respectively, and the corresponding rates for those without LDH determinations were 78% and 35%, respectively. The differences between these rates were not significant (P = .87). Similarly, the difference in survival between the two groups did not achieve statistical significance when evaluated for the entire population within each group (P = .41) and when evaluated by disease extent (P = .51). The median age of the study population was 60 years (range, 32 to 84). There were 204 men and 84 women. ECOG PS was evaluated as 0 in 20 patients, 1 in 191, 2 in 61, and 3 in 16 patients. One hundred forty-one patients presented with LD. One hundred forty-seven had ED with metastases to the brain in 16, liver in 76, contralateral lung in three, bone in 65, and bone marrow in 39 patients. Based on assessable cases, the overall response rate for the study population was 79%. Ninety-six patients (36%) achieved a CR, and 112 (43%) achieved a PR. Table 1 shows the distribution of LDH values by protocol and extent of disease.
Overall, 115 patients (40%) had a pretreatment level of LDH < 193 IU/L. Of the 173 patients (60%) with elevated (> 193 IU/L) levels of LDH, 79 (28%) had marginally elevated values (194 to 274 IU/L), whereas 94 (32%) had highly elevated values (2 275 IU/L). LDH was elevated in 65 of 141 patients (46%) with LD and in 108 of 147 patients (73%) with ED, a difference that was statistically significant (P = 2 x 10-6). In patients with ED, the presence of an abnormal LDH level (> 193 IU/L) was significantly correlated with metastatic liver involvement (P = .000011, r = .369) and bone marrow involvement (P = .000076, r = .339) but not with a positive bone scan (P = .352, r = .0818). Interestingly, a significant negative correlation between elevated levels of LDH and metastatic brain involvement (P = .00033, r = -. 329) was observed. Levels of LDH did not correlate with Table 1. LDH Levels by Study Protocol and Extent of Disease
those with (288 patients) LDH determinations
Limited
(the study population) were compared with re-
LDH (IU/L)
spect to their general characteristics, response to treatment, and overall survival to ensure that inadvertent selection bias did not influence the results. The two groups were comparable with respect to age, sex, extent of disease, and PS. The
CAV, TPN BR314 NOS Total
275
< 193
194-274
>275
17 7 13 2 39
18 1 5 2 26
26 4 0 9 39
18 9 1 12 40
44 7 4 13 68
Downloaded from ascopubs.org by University of Newcastle Upon Tyne on April 7, 2019 from 128.240.208.034 Copyright © 2019 American Society of Clinical Oncology. All rights reserved.
957
PROGNOSTIC SIGNIFICANCE OF LDH IN SCLC
the presence of pleural effusion (49 patients), or with hyponatremia (40 patients). Ninety-four percent of patients (108 of 115) with LD and 90% (156 of 173) of those with ED were assessable for response. Table 2 illustrates that the total response rate (CR + PR) of 89 of 108 (82%) patients with normal LDH was higher
100
a
60so 40 20
Z 0
than the rate of 119 of 156 (76%) patients with
elevated LDH, but this difference was not statistically significant (P = .28). The corresponding CR rates were 51 of 108 (47.2%) patients and 45 of 156 (29%) patients, respectively; this difference is statistically significant (P = .0028). When analyzed by extent of disease, LD patients with normal levels of LDH had a significantly higher CR rate, 44 of 76 (58%), when compared with those with elevated levels of LDH, 26 of 65 (40%) (P = .026). The response rates of patients with ED did not differ significantly between those with normal versus elevated levels of LDH (P = .30). Patients with elevated values of LDH had shorter survival times (log-rank analysis P = .0001). As illustrated in Fig 1, the median survival time for all patients with an elevated LDH level was 39 weeks, whereas for those with a normal LDH, it was 53 weeks. Superior 1- and 2-year survival rates of 54% and 16%, respectively, were demonstrated for those patients with normal levels of LDH relative to those patients with elevated levels of LDH, 33% and 6%, respectively (P = .00026). Of 28 long-term survivors (Ž 2 years) at the time of analysis, 17 (61%) had an initially normal LDH level. A negative correlation (log-rank analysis, P = .0001) between survival time and the quantitative level of LDH is demonstrated in Fig 2. The overall relative death risk for patients with ele-
80
L
rr
1
No. of Patients With LDH (IU/L) > 193
Total
CR PR Failure or no change Response not assessable Total Response rate
51 38 19 7 115 82%* (89/108) 47%t (51/108)
45 74 37 17 173 76% (119/156) 29% (45/156)
96 112 56 24 288 79% (208/264) 36% (96/264)
*P = .28 (two-tailed Fisher's exact test). tP = .0028 (two-tailed Fisher's exact test).
i
i
...........
t·
3
4
5
6
vated LDH when compared with those with normal values was 1.35:1. When analyzed by extent of disease a higher relative death rate was maintained in those with LD (1.63:1, P = .0015) and in those with ED (1.18:1); however, statistical significance was not achieved in the latter group (P = .434). The survival curves for patients grouped by extent of disease and by different ranges of LDH is illustrated in Fig 3. Poorer survival was apparent in LD patients with increasing levels of LDH. Interestingly, the median survival of LD patients with the highest range of LDH (> 275 IU/L) was not significantly different from that of all ED patients. The survival of patients with ED did not differ significantly according to the level of LDH. " 100 'i
rP
i'i
60
;
6
i·
40
20
Complete response rate
2
Fig 1. Survival curves from date of first treatment for patients with normal (_< 193 IU/L) initial LDH levels and elevated (> 193 IU/L) initial LDH levels. Survival was significantly better for the normal LDH group. (-) _ 193 IU/L, no. of patients 173, median survival 39 weeks; P < .0005 for both groups.
Table 2. Relationship of LDH Level and Response to Treatment
< 193
I
^
Time (Years)
S 80
Response Status
I
0
a 0
f
1
2
3
4
5
6
Time (Years) Fig 2. Survival curves from date of first treatment for patients grouped by different ranges of initial LDH values. (-) 0-193 IU/L, no. of patients 102, median survival 53 weeks; (....) 194-274 IU/L, no. of patients 110, median survival 40 weeks; (---) 275-999 IU/L, no. of patients 56, median survival 24 weeks; P < .0001 for all groups. A clear trend to poorer survival with increasing initial LDH values is apparent. This variation is statistically significant by the log-rank test.
Downloaded from ascopubs.org by University of Newcastle Upon Tyne on April 7, 2019 from 128.240.208.034 Copyright © 2019 American Society of Clinical Oncology. All rights reserved.
958
SAGMAN ET AL
100 a
so8 60
6
40 20
1
2
3
4
Time (Years)
Fig 3. Survival curves from date of first treatment for patients grouped by extent of disease and by different ranges of initial LDH levels. (-) LD, 0-193 IU/L, no. of patients 76, median survival 62 weeks, P < .0042; (..-.) LD, 194-274 lU/L, no. of patients 39, median survival 56 weeks, P < .0042; (---) LD, 275-999 IU/L, no. of patients 26, media in survival 37 weeks, P < .0042. The difference in mediain survival of patients in the latter group (LD, 275-999 IU/L) and that of all patients with ED (----; 0-999 IU/L, no. of patients 328, median survival 35 weeks) is not statistically significant, P < .3000.
To assess whether the initial LDH va lue contributed prognostic information not alrea dy obtained by other available variables, the surviv al data with an LDH threshold of 193 IU/L was adjusted by log-rank analysis for ECOG perform dance status, extent of disease, and type of study. A•s shown in ficant when Table 3, the P value remained signify adjustment was made for PS (P = .000:3), extent of disease (P = .0137), and both PS an d extent of disease (P = .0228). When adjusted fo,r these factors and for the type of clinical study the difference still remained significant (P = .0437). The analysis therefore suggests an indepe ndent prognostic value for LDH. DISCUSSION The present report suggests that LDH is a significant independent pretreatment prognostic factor in patients with SCLC. The resuilts indicate that pretreatment serum levels of LD H correlate with objective response to treatmen t and with survival. The independent value of LDH was Table 3. Adjustment of Survival for Different Factors in the Presence of an Abnormal LIDH
No adjustment Adjustment for PS Extent of disease Type of study PS + extent of disease PS + extent of disease + type of study *Wilcoxon-Gehan X' test for trend.
Level .0001 .0003 .0137
.0002 .0228
.0437
confirmed by persistence of statistical significance following adjustment for the most predictive prognostic parameters in SCLC to date-extent of disease and PS-as well as following adjustment for the different treatment protocols. It has been recognized that a number of other serum markers have prognostic implications in SCLC. Abnormal levels of serum sodium, albumin, alkaline phosphatase, B-glutamyl transpeptidase, calcitonin, CEA, arginine-vasopressin, and 67 ' ,19-22 NSE can all influence prognosis adversely. Correlations between an elevated marker and involvement of a specific disease site have been suggested,' though no such relationships have been proven with any marker. Similarly, in our study, levels of LDH were positively correlated with both liver and bone marrow metastasis; hence, this marker cannot be viewed as site-specific. It would be interesting in this respect to examine the relationship of LDH isoenzymes to specific metastatic sites of disease. The negatively correlated relationship between serum LDH levels and brain involvement is intriguing and may relate to factors governed by the blood-brain barrier. Assessment of LDH levels obtained concurrently from serum and CSF may be of interest in this regard. In this study, abnormal levels of LDH correlated with more ED. Moreover, elevated levels of LDH in patients with LD predicted for poorer survival relative to those with LDH in the normal range. In fact, our study demonstrated that the survival of a subset of LD patients-those with markedly elevated levels of LDH ( 2 275 IU/L)was not significantly different from that of patients with ED. A possible explanation of these relationships may be that LDH estimates tumor burden in patients with LD and suggests that LDH identifies occult disease not detected by current staging investigations. Our finding that the survival of patients with ED did not correlate with the level of serum LDH is in contrast with recent reports,23-25 which demonstrate a prognostic value for serum LDH in patients with ED. This dichotomy may reflect the choice of our study population, which may have been restricted to a more favorable cohort by the exclusion of patients with poorer PS in our studies. We suggest, therefore, that the prognostic significance of serum LDH be further confirmed, particularly in studies that accommodate the full spectrum of the SCLC population. The finding that serum LDH correlates with
Downloaded from ascopubs.org by University of Newcastle Upon Tyne on April 7, 2019 from 128.240.208.034 Copyright © 2019 American Society of Clinical Oncology. All rights reserved.
959
PROGNOSTIC SIGNIFICANCE OF LDH IN SCLC
disease extent as demonstrated by this report and by others may bear on current staging practices in SCLC. It seems probable that patients with SCLC present with a continuous spectrum of disease extent and that prognosis is related to this. By implication, the categories of LD and ED must be regarded as only a crude measure of tumor mass. For example, several studies indicate that the primary tumor size is correlated with survival and that the number of metastatic sites is also an important prognostic indicator.26 In agreement with our findings, Byhardt et a127 used serum LDH to stratify patients according to tumor mass using a modified tumor-node-metastasis classification. The four prognostic groups identified in their study demonstrated the limitation of the current LD-ED staging system in the assessment of response to treatment and the prognosis of survival. Speculation on the biologic significance of LDH in malignancy may be inferred from the present study. As a glycolytic enzyme and an end product of the Embden-Meyerhof pathway, LDH is ubiquitous and associated with a multitude of pathologic processes. Elevated levels of LDH have been documented in association with enhanced metabolic activity (ie, infections) and as a sequela to tissue injury (ie, pulmonary and myocardial infarction). In neoplasia, elevation of LDH in serum may reflect the release of the enzyme secondary to inflammation or tumor necrosis, higher levels correlating with increased tumor mass. Furthermore, the demonstration that higher intracellular levels of LDH are present in high- versus lowgrade lymphoma28 may reflect altered cell kinetics pertaining to the biologic behavior of a tumor. In the context of the present study, unfavorable prognosis in those with elevated levels of LDH may not only reflect tumor mass but also tumor aggressivity. LDH therefore may serve as a measure of overall disease activity. In fact, the use of three biochemical markers, including LDH by Ganz et a12 9 to serially monitor response to therapy in patients with SCLC, showed a significant direct correlation between disease activity and serum LDH. In contradistinction to the present report and to at least six other reported studies23'2 5 ' 30 -32 in which elevated levels of serum LDH predict for poor outcome in patients with SCLC, Hainsworth et a133 have reported that elevated serum level at the time of diagnosis was predictive of improved survival. This finding is perplexing in that elevated
serum levels of LDH have been associated with poor prognosis in other types of neoplasms.8-12 A possible explanation, nevertheless, may be that other prognostic attributes not included in the study of Hainsworth et al may dictate prognosis in addition to LDH levels and that interactions of such variables with LDH may reflect prognosis more accurately. Hainsworth et al, in fact, state that it is conceivable that such prognostic factors in their patient population were possibly overlooked or were not possible to analyze due to the small number of patients in the study. In this regard, recent advances in the analysis of prognostic factors have focused on the use of multivariate models, which endeavor to demonstrate that several factors in addition to disease extent provide separate and additional value in the prediction of prognosis. In this respect, the correlation of serum markers with other predictors of prognosis is particularly noteworthy. Ohnoshi et a131 illustrated the value of assessing serum LDH and albumin in combination with conventional prognostic factors in a multivariate analysis. We have previously reported on the combined use of LDH and CEA as prognostic factors for SCLC.32 Both CEA and LDH levels are independent prognostic factors in SCLC; however, in comparison to LDH, which stratifies patients into three prognostic groups, the relevance of CEA was restricted to a subset of patients with LD. Interactions between LDH and other prognostic factors were used by Osterlind and Anderson, 23 Cerney et al, 24 and Albain et a125 to define prognostic risk groups by multivariate analysis. All these studies corroborated the inclusion of LDH as a valuable prognostic factor in patients with SCLC. One particular advantage of LDH is that this laboratory test is comparatively inexpensive and readily available. In this regard, Dole34 reported on a perfect correlation between the absence of bone marrow involvement and normal levels of serum LDH and recommended that the routine use of bone marrow evaluation in such patients be omitted. Similarly, in this report, only 1.7% (two of 110) of patients with normal LDH levels were found to have detectable bone marrow involvement. The corresponding rates in other series have been somewhat higher at 10% to 16%.35.36 Never-
theless, it is conceivable that in conjunction with other routinely performed laboratory parameters, the assessment of LDH may define disease burden
Downloaded from ascopubs.org by University of Newcastle Upon Tyne on April 7, 2019 from 128.240.208.034 Copyright © 2019 American Society of Clinical Oncology. All rights reserved.
960
SAGMAN ET AL
and obviate the necessity of more complex and invasive staging investigations. Consequently, the requirements for the assessment of prognosis and treatment planning could be simplified, patient discomfort minimized, and health care costs might be reduced. In conclusion, LDH is a useful independent pretreatment prognostic attribute in patients with
SCLC that correlates with extent of disease and response to treatment and survival following diagnosis. It is suggested that LDH complements other prognostic factors and, hence, may contribute to a more accurate assessment of patient status. Finally, the role of LDH in stratifying patients into treatment groups during the conduct of clinical trials should be further studied.
REFERENCES 1. Minna JD, Higgins GA, Glatstein EI: Cancer of the lung, in DeVita VT, Hellman S, Rosenberg SA (eds): Cancer, Principles and Practice of Oncology. Philadelphia, PA, Lippincott, 1985, pp 507-598 2. Morstyn G, lhde DC, Lichter AS, et al: Small cell lung cancer 1973-1983: Early progress and recent obstacles. Int J Radiat Oncol Biol Phys 10:515-539, 1984 3. Cohen MH, Ihde DC, Bunn, et al: Cyclic alternating combination chemotherapy for small cell bronchogenic carcinoma. Cancer Treat Rep 63:163-170, 1979 4. Feld R, Ginsberg RJ, Payne DG: Treatment of small cell lung cancer, in Roth JA, Ruckdeschel JC, Weisenburger TH (eds): Thoracic Oncology. Philadelphia, PA, Saunders, 1989, pp 229-262 5. Stanley KE: Prognostic factors in lung cancer, in Aisner J (ed): Lung Cancer. New York, NY, Churchill Livingstone, 1985, pp 41-66 6. Sculier JP, Feld R, Evans WK, et al: Carcinoembryonic antigen: A useful prognostic marker in small cell lung cancer. J Clin Oncol 3:1349-1354, 1985 7. Carney DN, Ihde DC, Cohen MH, et al: Serum neuron specific enolase: A marker for disease extent and response to therapy of small cell carcinoma. Lancet 1:583-585, 1982 8. Schilling RF, McKnight B, Crowley JJ: Prognostic value of serum lactate dehydrogenase level in Hodgkin's disease. J Lab Clin Med 99:382-387, 1982 9. Hagberg H, Siegbahn A: Prognostic value of serum lactic dehydrogenase in non-Hodgkin's lymphoma. Scand J Haematol 31:49-65, 1983 10. Brereton HD, Simon R, Pomeroy TC: Pretreatment lactate dehydrogenase predicting metastatic spread in Ewing's sarcoma. Ann Intern Med 83:352-354, 1975 11. Simonsson B, Grenning G, Kallander C, et al: Prognostic value of serum lactic dehydrogenase (S-LDH) in multiple myeloma. Eur J Clin Oncol 17:336-339, 1987 12. Barlogie B, Smallwood L, Smith T, et al: High serum levels of lactic dehydrogenase identify a high-grade lymphoma-like myeloma. Ann Intern Med 110:521-525, 1989 13. Feld R, Pringle JF, Evans WK, et al: Combined modality treatment of small cell carcinoma of the lung. Arch Intern Med 141:469-473, 1981 14. Clamon GH, Feld R, Evans WK, et al: The effect of adjuvant central IV hyperalimentation on the survival and response to treatment of patients with small cell lung cancer: A randomized trial. Cancer Treat Rep 69:167-177, 1985 15. Evans WK, Feld R, Murray N, et al: Superiority of alternating non-cross-resistant chemotherapy in extensive small cell lung cancer: The results of a multicentre random-
ized National Cancer Institute of Canada clinical trial. Ann Intern Med 107:451-458, 1987 16. Feld R, Evans WK, Coy P, et al: Canadian multicentre randomized trial comparing sequential and alternating administration of two non-cross-resistant chemotherapy combinations in patients with limited small cell carcinoma of the lung. J Clin Oncol 5:1401-1409, 1987 17. Kaplan EL, Meier P: Non parametric estimation from incomplete observations. Am Stat Assoc J 53:457-481, 1958 18. Peto R, Pike MC, Armitage P, et al: Design and analysis of randomized trials requiring prolonged observations of each patient II. Analysis and examples. Br J Cancer 35:1-39, 1977 19. Souhami RL, Bradbury J, Geddes DM, et al: Prognostic significance of laboratory parameters measured at diagnosis in small cell carcinoma of the lung. Cancer Res 45:2878-2882, 1985 20. Cohen MH, Makuch R, Johnston-Early A, et al: Laboratory parameters as an alternative to performance status in prognostic stratification of patients with small cell lung cancer. Cancer Treat Rep 65:187-195, 1981 21. Maurer LH, Pajak TF: Prognostic factors in small cell carcinoma of the lung: A Cancer and Leukemia Group B study. Cancer Treat Rep 65:767-774, 1981 22. Wallach SR, Royston I, Taetle R, et al: Plasma calcitonin as marker of disease activity in patients with small cell carcinoma of the lung. J Clin Endocrin Metab 53:602606, 1981 23. Osterlind K, Andersen PK: Prognostic factors in small cell lung cancer: Multivariate model based on 778 patients treated with chemotherapy with or without irradiation. Cancer Res 46:4189-4194, 1986 24. Cerney T, Blair V, Anderson H, et al: Pretreatment prognostic factors and scoring sytem in 407 small cell lung cancer patients. Int J Cancer 39:146-149, 1987 25. Albain KS, Crowley JJ, Le Blanc M, et al: Determinants of improved outcome in small-cell lung cancer: An analysis of the 2,580-patient Southwest Oncology Group data base. J Clin Oncol 8:563-1574, 1990 26. lhde DC, Makuch RW, Carney DN, et al: Prognostic implications of stage of disease and sites of metastasis in patients with small cell carcinoma of the lung treated with intensive combination chemotherapy. Am Rev Respir Dis 123:500-507, 1981 27. Byhardt RW, Hartz A, Libnoch JA, et al: Prognostic influence of TNM staging and LDH levels in small cell carcinoma of the lung (SCCL). Int J Radiat Oncol Biol Phys 12:771-777, 1986
Downloaded from ascopubs.org by University of Newcastle Upon Tyne on April 7, 2019 from 128.240.208.034 Copyright © 2019 American Society of Clinical Oncology. All rights reserved.
961
PROGNOSTIC SIGNIFICANCE OF LDH IN SCLC 28. Vezzoni MA, Lucchini R, Giardini R, et al: Lactate dehydrogenase levels in cellular extracts of human malignant lymphomas. Tumori 69:279-282, 1983 29. Ganz PA, Ma PY, Wang HJ, et al: Evaluation of three biochemical markers for serially monitoring the therapy of small cell lung cancer. J Clin Oncol 5:472-479, 1987 30. Vincent MD, Ashley SE, Smith IA: Prognostic factors in small cell lung cancer: A simple prognostic index is better than conventional staging. Eur J Cancer Clin Oncol 23:1589-1599, 1987 31. Ohnoshi T, Hiraki S, Nakata Y, et al: Pretreatment serum albumin concentration and lactic dehydrogenase activity as prognostic factors in patients with small cell lung cancer. Acta Med Okayama 36:487-490, 1982 32. Wang T, Feld R, Sagman U, et al: The combined use
of LDH and CEA as independent prognostic markers for small cell lung cancer. Proc Am Assoc Cancer Res 27:162, 1986 (abstr 641) 33. Hainsworth JD, Johnson DH, Farese RV: Limited stage small cell lung cancer: Analysis of clinical prognostic factors. South Med J 80:1518-1522, 1987 34. Dole DC: Serum lactate dehydrogenase and bone marrow involvement in small cell carcinoma of the lung. N Engl J Med 712:1262, 1985 35. Osterlind K, Hirsch FR, Pedersen AG: Serum lactate dehydrogenase and bone marrow metastasis in small cell carcinoma of the lung. N Engl J Med 313:759, 1985 36. Johnson DH, Greco AF: Serum lactate dehydrogenase and bone marrow metastasis in small cell carcinoma of the lung. N Engl J Med 313:759, 1985
Downloaded from ascopubs.org by University of Newcastle Upon Tyne on April 7, 2019 from 128.240.208.034 Copyright © 2019 American Society of Clinical Oncology. All rights reserved.
rates for patients with elevated LDH were 39 weeks and 33%and 6%, respectively, whereas for those with a normal LDH level these were 53 weeks and 54% and 16%, respectively. Patients with LD and elevated levels of LDH manifested a higher relative death rate (1.63:1)
THE
measure of individual outcome as well as optimize patient stratification into clinical trials and therefore lead to improved assessment of such studies. The most important prognostic factors identified to date have included extent of disease and clinical performance status (PS). 5 The prognostic importance of several biologic markers has also been assessed in patients with SCLC. To date, most substances examined have lacked stringent sensitivity and specificity as diagnostic adjuncts. Pretreatment serum levels of biologic markers, such as the carcinoembryonic antigen (CEA) and neuron-specific enolase (NSE), however, are important determinants of prognosis6 ' 7 and hence may have a role in the design of treatment plans used to treat patients with SCLC. The possible role of the serum enzyme lactate dehydrogenase (LDH) as a prognostic attribute has recently been examined in patients with lymphoma,' 9 selected histologic types of sarcoma,1 0 and multiple myeloma., 1 ' 2 We have measured the level of serum LDH in patients with SCLC before treatment on clinical studies at our institutions. We report our analysis of the value of LDH in the staging and prognosis of patients with SCLC.
SURVIVAL of patients with small-cell carcinoma of the lung (SCLC) has improved with the advent of new chemotherapeutic agents and the use of combination chemotherapy.' Without treatment, the median length of survival after diagnosis is 1 to 3 months.2 Combination chemotherapy produces objective tumor regression in as many as 90% of patients and median survival times of approximately 8 to 11 months for patients with extensive disease (ED) and more than a year for those with limited disease (LD).3 4 With this improved outcome, it is desirable to identify prospectively patients with a more favorable prognosis. This knowledge may provide a more accurate
From The Ontario Cancer Institute Incorporating The Princess Margaret Hospital, Toronto General Hospital, Mt Sinai Hospital, Sunnybrook Medical Center, University of Toronto, Toronto; Ontario Cancer Treatment and Research Foundation; and Ottawa Regional Cancer Centre, Ottawa, Canada. SubmittedApril 9, 1990; accepted December 26, 1990. Address reprintrequests to Ronald Feld, MD, The Ontario Cancer Institute, 500 Sherboume St, Toronto, Ontario M4X 1K9 Canada. 0 1991 by American Society of Clinical Oncology. 0732-183X91/0906-0006$3.00o/0
954
when compared with patients with LD and LDH in the
normal range (P = .0083). The survival of patients with ED did not differ between those with normal and elevated levels of LDH (P = .273). A significant survival advantage persisted for patients with LDH in the normal range following adjustments for extent of disease, performance status (PS), and treatment protocol (P = .044, log-rank analysis). In conclusion, serum LDH appears to be a significant independent pretreatment prognostic factor in patients with SCLC that correlates with stage of disease, response to treatment, and survival. J Clin Oncol 9:954-961. o 1991 by American Society of ClinicalOncology.
Journalof Clinical Oncology, Vol 9, No 6 (June), 1991: pp 954-961
Downloaded from ascopubs.org by University of Newcastle Upon Tyne on April 7, 2019 from 128.240.208.034 Copyright © 2019 American Society of Clinical Oncology. All rights reserved.
955
PROGNOSTIC SIGNIFICANCE OF LDH IN SCLC MATERIALS AND METHODS All patients with SCLC managed at University of Toronto affiliated hospitals from December 1976 to December 1986 were considered for retrospective analysis in this study. All patients had histologic or cytologic proof of SCLC, had received no prior treatment with chemotherapy, radiotherapy, or surgery, and had a performance status (PS) from 0 to 3 on the Eastern Cooperative Oncology Group (ECOG) scale. They were excluded if they had a history of previous malignancy, unless disease-free for at least 5 years, or if the tumor was a nonmelanoma skin cancer. Patients were free of active cardiac disease, severe hepatic dysfunction (serum bilirubin > 2 mgi100 mL), and had normal renal and hematologic function. Those patients included in a study of total parenteral nutrition as an adjunct to chemotherapy required (see below) assessment for evaluation of disease and freedom from CNS metastases, serious uncontrolled infection, superior vena cava obstruction, hyponatremia (< 125 mEq/L), or severe hypoalbuminemia (< 2.5 g/100 mL). Patients on all other studies did not have these exclusions. All patients had a complete history, physical examination, evaluation of performance status, and investigations to define the extent of their disease, which included routine hematology and chemistry, chest x-ray, bone, liver, and spleen radioisotope scans, and bone marrow aspirate and biopsy. Patients with CNS symptoms and most other patients had pretreatment brain scintiscans or computed tomographic scans. Those patients with clinical evidence of tumor confined to a hemithorax and ipsilateral supraclavicular nodes (inclusive of cytology-negative pleural effusions) were considered to have LD. All others with evidence of disease spread beyond the extent defined for LD were designated to have ED. Pretreatment serum LDH determinations were performed on 288 of 614 patients who were managed during the 10-year study interval. Serum LDH determinations were not part of the routine pretreatment investigations, but were performed randomly at the discretion of attending physicians as part of a battery of tests aimed at evaluating hepatic dysfunction specifically and overall disease activity in general. No other tests were performed as a direct result of elevated LDH levels during the pretreatment interval, and therefore, the staging investigations of the study population were not modified. The proportion of patients on whom LDH levels were determined as a function of year of study was relatively uniform throughout the period covered by the study. The two hundred eighty-eight patients on whom serum LDH levels were evaluated form the basis of this report. Two hundred forty of these patients were managed in controlled clinical trials at our institutions along with other patients on whom LDH levels were not determined. The remaining 48 patients were not included in a specific trial but had identical investigations and staging procedures. One hundred fifty-one patients were derived from previously reported study" in which initial treatment consisted of 2 cyclophosphamide 900 mg/m , doxorubicin (Adriamycin; Adria Laboratories, Columbus, OH) 45 mg/m2, and vincristine 2 mg by bolus intravenous (IV) injection (CAV).
Treatments were repeated at 3-week intervals for a total of six courses unless tumor progression was evident. Approximately 1 week following the third course of chemotherapy, the responding patient underwent whole-brain prophylactic irradiation (2,000 cGy in five equal fractions). Following six cycles of CAV, the pretreatment volume of intrathoracic tumor was irradiated in responding patients with both LD and ED (2,500 cGy in 10 equal fractions over 2 weeks). At relapse, re-treatment with the CAV regimen was generally attempted, and if reinduction failed, further treatment was administered at the discretion of the attending physician. 4 The 43 patients, derived from a second published trial,' were treated in a randomized fashion to evaluate the efficacy of total parenteral nutrition (TPN) as an adjunct to chemotherapy (part of a multicenter National Cancer Institute-sponsored trial). In this study, patients received two courses of CAV, which consisted of cyclophosphamide 2 1,200 mg/m 2, Adriamycin 70 mg/m , and vincristine 2 mg administered IV by bolus injection at 3-week intervals. Those achieving an objective response received two further courses of lower dose CAV followed by prophylactic cranial irradiation (2,000 cGy in five equal fractions), and in those with LD, chest irradiation was administered (2,500 cGy in 10 equal fractions over a period of 2 weeks). Patients were then treated with etoposide (VP-16) 100 mg/m2 for 3 days and methotrexate 50 mg/m2 IV on day 1 with leucovorin (VMC) every 3 weeks for five cycles. Treatment with further 2 cycles of CAV to a total Adriamycin dose of 450 mg/m was followed by five additional courses of VMC. Patients who did not respond to the two initial courses of CAV were treated with VMC to disease progression. Patients randomized to receive TPN in the second protocol were started on nutritional support 7 days before the initial course of chemotherapy, and TPN was discontinued 48 hours following the administration of the second cycle of CAV. A third group of patients were obtained from two trials, (combined here as BR), which evaluated the efficacy of two alternating noncross-resistant chemotherapeutic regimens 5 in the management of patients with ED, BR ,'4 and LD, 6 BR,.' In the BR 4 trial, 46 patients were randomized to receive either six courses of IV cyclophosphamide 1,000 2 mg/m 2, doxorubicin 50 mg/m , and vincristine 2 mg (CAV) at 3-week intervals, or CAV alternating with VP-16 100 mg/m2 2 days 1 to 3 and cisplatin 25 mg/m on days 1 to 3 every 3 who progressed on management of patients weeks. Further the alternating regimen was instituted at the discretion of the physician. The four patients derived from the BR 3 trial were randomized to receive three courses of CAV followed by three courses of VP-16 and cisplatin for a total of six courses. All responders received prophylactic cranial irradiation (2,000 cGy in five equal fractions) following three cycles of treatment and were randomized for locoregional control with either standard (2,500 cGy) or high-dose (3,750 cGy) irradiation regimens. Forty-eight patients were not on study (NOS) protocols, and comprised those not meeting eligibility criteria for concurrently active trials. The characteristics of the 48 patients including age, sex distribution, performance status, and disease extent were similar to those of the patients who did enter study protocols. Moreover, as a group, their response to treatment and overall survival was not signifi-
Downloaded from ascopubs.org by University of Newcastle Upon Tyne on April 7, 2019 from 128.240.208.034 Copyright © 2019 American Society of Clinical Oncology. All rights reserved.
956
SAGMAN ET AL
cantly different from those who were managed according to clinical trial protocols. Although management of such patients was based on the investigator's option, most received a chemotherapeutic regimen that included cyclophosphamide, Adriamycin, and vincristine (a modified CAV protocol). Criteria for response to treatment were defined uniformly in all patients in this study. Complete response (CR) was defined as disappearance of all known neoplastic disease. A partial response (PR) was defined as a decrease > 50 in the product of the two greatest perpendicular diameters of well-outlined lesions, lasting for at least 3 weeks in the absence of progressive disease elsewhere or the development of any new lesions. In the assessment of poorly defined lesions, an estimated 75% decrease in the cross-sectional area was required to qualify for a PR, again in the absence of progressive disease elsewhere or the development of any new lesions. Progression of disease was defined as a greater than 25% increase in the cross-sectional area of one or more lesions or the occurrence of any new lesions. Survival time was measured from the date of the first chemotherapy administration to the date of death. LDH Determinationand Normal Range Pretreatment serum LDH activity was assayed by the spectrophotometric absorbency of nicotinamide-adenine dinucleotide at 340 nm, following catalytic oxidation of lactate to pyruvate. The normal reference range based on 276 apparently healthy subjects was considered to be between 100 to 193 IU/L (ACA, Du Pont automatic clinical analyzer). No isoenzyme fractionation was undertaken. LDH assay determinations were performed in the respective departments of Clinical Biochemistry at The Princess Margaret Hospital and the Toronto General Hospital (Toronto). StatisticalConsiderations Survival curves were calculated by the method of Kaplan and Meier.17 The log-rank method"' was used to test the statistical significance of the difference between survival curves, as well as to do comparisons adjusted for other prognostic factors. P values for differences between proportions were calculated with Fisher's exact test (two-tailed). Correlations between LDH level (normal or elevated) and specific metastatic sites were assessed with the product moment correlation coefficient (r). RESULTS Of the total 614 patients managed during the study interval, those without (326 patients) and
overall response and CR rates for those with LDH determination were 79% and 36%, respectively, and the corresponding rates for those without LDH determinations were 78% and 35%, respectively. The differences between these rates were not significant (P = .87). Similarly, the difference in survival between the two groups did not achieve statistical significance when evaluated for the entire population within each group (P = .41) and when evaluated by disease extent (P = .51). The median age of the study population was 60 years (range, 32 to 84). There were 204 men and 84 women. ECOG PS was evaluated as 0 in 20 patients, 1 in 191, 2 in 61, and 3 in 16 patients. One hundred forty-one patients presented with LD. One hundred forty-seven had ED with metastases to the brain in 16, liver in 76, contralateral lung in three, bone in 65, and bone marrow in 39 patients. Based on assessable cases, the overall response rate for the study population was 79%. Ninety-six patients (36%) achieved a CR, and 112 (43%) achieved a PR. Table 1 shows the distribution of LDH values by protocol and extent of disease.
Overall, 115 patients (40%) had a pretreatment level of LDH < 193 IU/L. Of the 173 patients (60%) with elevated (> 193 IU/L) levels of LDH, 79 (28%) had marginally elevated values (194 to 274 IU/L), whereas 94 (32%) had highly elevated values (2 275 IU/L). LDH was elevated in 65 of 141 patients (46%) with LD and in 108 of 147 patients (73%) with ED, a difference that was statistically significant (P = 2 x 10-6). In patients with ED, the presence of an abnormal LDH level (> 193 IU/L) was significantly correlated with metastatic liver involvement (P = .000011, r = .369) and bone marrow involvement (P = .000076, r = .339) but not with a positive bone scan (P = .352, r = .0818). Interestingly, a significant negative correlation between elevated levels of LDH and metastatic brain involvement (P = .00033, r = -. 329) was observed. Levels of LDH did not correlate with Table 1. LDH Levels by Study Protocol and Extent of Disease
those with (288 patients) LDH determinations
Limited
(the study population) were compared with re-
LDH (IU/L)
spect to their general characteristics, response to treatment, and overall survival to ensure that inadvertent selection bias did not influence the results. The two groups were comparable with respect to age, sex, extent of disease, and PS. The
CAV, TPN BR314 NOS Total
275
< 193
194-274
>275
17 7 13 2 39
18 1 5 2 26
26 4 0 9 39
18 9 1 12 40
44 7 4 13 68
Downloaded from ascopubs.org by University of Newcastle Upon Tyne on April 7, 2019 from 128.240.208.034 Copyright © 2019 American Society of Clinical Oncology. All rights reserved.
957
PROGNOSTIC SIGNIFICANCE OF LDH IN SCLC
the presence of pleural effusion (49 patients), or with hyponatremia (40 patients). Ninety-four percent of patients (108 of 115) with LD and 90% (156 of 173) of those with ED were assessable for response. Table 2 illustrates that the total response rate (CR + PR) of 89 of 108 (82%) patients with normal LDH was higher
100
a
60so 40 20
Z 0
than the rate of 119 of 156 (76%) patients with
elevated LDH, but this difference was not statistically significant (P = .28). The corresponding CR rates were 51 of 108 (47.2%) patients and 45 of 156 (29%) patients, respectively; this difference is statistically significant (P = .0028). When analyzed by extent of disease, LD patients with normal levels of LDH had a significantly higher CR rate, 44 of 76 (58%), when compared with those with elevated levels of LDH, 26 of 65 (40%) (P = .026). The response rates of patients with ED did not differ significantly between those with normal versus elevated levels of LDH (P = .30). Patients with elevated values of LDH had shorter survival times (log-rank analysis P = .0001). As illustrated in Fig 1, the median survival time for all patients with an elevated LDH level was 39 weeks, whereas for those with a normal LDH, it was 53 weeks. Superior 1- and 2-year survival rates of 54% and 16%, respectively, were demonstrated for those patients with normal levels of LDH relative to those patients with elevated levels of LDH, 33% and 6%, respectively (P = .00026). Of 28 long-term survivors (Ž 2 years) at the time of analysis, 17 (61%) had an initially normal LDH level. A negative correlation (log-rank analysis, P = .0001) between survival time and the quantitative level of LDH is demonstrated in Fig 2. The overall relative death risk for patients with ele-
80
L
rr
1
No. of Patients With LDH (IU/L) > 193
Total
CR PR Failure or no change Response not assessable Total Response rate
51 38 19 7 115 82%* (89/108) 47%t (51/108)
45 74 37 17 173 76% (119/156) 29% (45/156)
96 112 56 24 288 79% (208/264) 36% (96/264)
*P = .28 (two-tailed Fisher's exact test). tP = .0028 (two-tailed Fisher's exact test).
i
i
...........
t·
3
4
5
6
vated LDH when compared with those with normal values was 1.35:1. When analyzed by extent of disease a higher relative death rate was maintained in those with LD (1.63:1, P = .0015) and in those with ED (1.18:1); however, statistical significance was not achieved in the latter group (P = .434). The survival curves for patients grouped by extent of disease and by different ranges of LDH is illustrated in Fig 3. Poorer survival was apparent in LD patients with increasing levels of LDH. Interestingly, the median survival of LD patients with the highest range of LDH (> 275 IU/L) was not significantly different from that of all ED patients. The survival of patients with ED did not differ significantly according to the level of LDH. " 100 'i
rP
i'i
60
;
6
i·
40
20
Complete response rate
2
Fig 1. Survival curves from date of first treatment for patients with normal (_< 193 IU/L) initial LDH levels and elevated (> 193 IU/L) initial LDH levels. Survival was significantly better for the normal LDH group. (-) _ 193 IU/L, no. of patients 173, median survival 39 weeks; P < .0005 for both groups.
Table 2. Relationship of LDH Level and Response to Treatment
< 193
I
^
Time (Years)
S 80
Response Status
I
0
a 0
f
1
2
3
4
5
6
Time (Years) Fig 2. Survival curves from date of first treatment for patients grouped by different ranges of initial LDH values. (-) 0-193 IU/L, no. of patients 102, median survival 53 weeks; (....) 194-274 IU/L, no. of patients 110, median survival 40 weeks; (---) 275-999 IU/L, no. of patients 56, median survival 24 weeks; P < .0001 for all groups. A clear trend to poorer survival with increasing initial LDH values is apparent. This variation is statistically significant by the log-rank test.
Downloaded from ascopubs.org by University of Newcastle Upon Tyne on April 7, 2019 from 128.240.208.034 Copyright © 2019 American Society of Clinical Oncology. All rights reserved.
958
SAGMAN ET AL
100 a
so8 60
6
40 20
1
2
3
4
Time (Years)
Fig 3. Survival curves from date of first treatment for patients grouped by extent of disease and by different ranges of initial LDH levels. (-) LD, 0-193 IU/L, no. of patients 76, median survival 62 weeks, P < .0042; (..-.) LD, 194-274 lU/L, no. of patients 39, median survival 56 weeks, P < .0042; (---) LD, 275-999 IU/L, no. of patients 26, media in survival 37 weeks, P < .0042. The difference in mediain survival of patients in the latter group (LD, 275-999 IU/L) and that of all patients with ED (----; 0-999 IU/L, no. of patients 328, median survival 35 weeks) is not statistically significant, P < .3000.
To assess whether the initial LDH va lue contributed prognostic information not alrea dy obtained by other available variables, the surviv al data with an LDH threshold of 193 IU/L was adjusted by log-rank analysis for ECOG perform dance status, extent of disease, and type of study. A•s shown in ficant when Table 3, the P value remained signify adjustment was made for PS (P = .000:3), extent of disease (P = .0137), and both PS an d extent of disease (P = .0228). When adjusted fo,r these factors and for the type of clinical study the difference still remained significant (P = .0437). The analysis therefore suggests an indepe ndent prognostic value for LDH. DISCUSSION The present report suggests that LDH is a significant independent pretreatment prognostic factor in patients with SCLC. The resuilts indicate that pretreatment serum levels of LD H correlate with objective response to treatmen t and with survival. The independent value of LDH was Table 3. Adjustment of Survival for Different Factors in the Presence of an Abnormal LIDH
No adjustment Adjustment for PS Extent of disease Type of study PS + extent of disease PS + extent of disease + type of study *Wilcoxon-Gehan X' test for trend.
Level .0001 .0003 .0137
.0002 .0228
.0437
confirmed by persistence of statistical significance following adjustment for the most predictive prognostic parameters in SCLC to date-extent of disease and PS-as well as following adjustment for the different treatment protocols. It has been recognized that a number of other serum markers have prognostic implications in SCLC. Abnormal levels of serum sodium, albumin, alkaline phosphatase, B-glutamyl transpeptidase, calcitonin, CEA, arginine-vasopressin, and 67 ' ,19-22 NSE can all influence prognosis adversely. Correlations between an elevated marker and involvement of a specific disease site have been suggested,' though no such relationships have been proven with any marker. Similarly, in our study, levels of LDH were positively correlated with both liver and bone marrow metastasis; hence, this marker cannot be viewed as site-specific. It would be interesting in this respect to examine the relationship of LDH isoenzymes to specific metastatic sites of disease. The negatively correlated relationship between serum LDH levels and brain involvement is intriguing and may relate to factors governed by the blood-brain barrier. Assessment of LDH levels obtained concurrently from serum and CSF may be of interest in this regard. In this study, abnormal levels of LDH correlated with more ED. Moreover, elevated levels of LDH in patients with LD predicted for poorer survival relative to those with LDH in the normal range. In fact, our study demonstrated that the survival of a subset of LD patients-those with markedly elevated levels of LDH ( 2 275 IU/L)was not significantly different from that of patients with ED. A possible explanation of these relationships may be that LDH estimates tumor burden in patients with LD and suggests that LDH identifies occult disease not detected by current staging investigations. Our finding that the survival of patients with ED did not correlate with the level of serum LDH is in contrast with recent reports,23-25 which demonstrate a prognostic value for serum LDH in patients with ED. This dichotomy may reflect the choice of our study population, which may have been restricted to a more favorable cohort by the exclusion of patients with poorer PS in our studies. We suggest, therefore, that the prognostic significance of serum LDH be further confirmed, particularly in studies that accommodate the full spectrum of the SCLC population. The finding that serum LDH correlates with
Downloaded from ascopubs.org by University of Newcastle Upon Tyne on April 7, 2019 from 128.240.208.034 Copyright © 2019 American Society of Clinical Oncology. All rights reserved.
959
PROGNOSTIC SIGNIFICANCE OF LDH IN SCLC
disease extent as demonstrated by this report and by others may bear on current staging practices in SCLC. It seems probable that patients with SCLC present with a continuous spectrum of disease extent and that prognosis is related to this. By implication, the categories of LD and ED must be regarded as only a crude measure of tumor mass. For example, several studies indicate that the primary tumor size is correlated with survival and that the number of metastatic sites is also an important prognostic indicator.26 In agreement with our findings, Byhardt et a127 used serum LDH to stratify patients according to tumor mass using a modified tumor-node-metastasis classification. The four prognostic groups identified in their study demonstrated the limitation of the current LD-ED staging system in the assessment of response to treatment and the prognosis of survival. Speculation on the biologic significance of LDH in malignancy may be inferred from the present study. As a glycolytic enzyme and an end product of the Embden-Meyerhof pathway, LDH is ubiquitous and associated with a multitude of pathologic processes. Elevated levels of LDH have been documented in association with enhanced metabolic activity (ie, infections) and as a sequela to tissue injury (ie, pulmonary and myocardial infarction). In neoplasia, elevation of LDH in serum may reflect the release of the enzyme secondary to inflammation or tumor necrosis, higher levels correlating with increased tumor mass. Furthermore, the demonstration that higher intracellular levels of LDH are present in high- versus lowgrade lymphoma28 may reflect altered cell kinetics pertaining to the biologic behavior of a tumor. In the context of the present study, unfavorable prognosis in those with elevated levels of LDH may not only reflect tumor mass but also tumor aggressivity. LDH therefore may serve as a measure of overall disease activity. In fact, the use of three biochemical markers, including LDH by Ganz et a12 9 to serially monitor response to therapy in patients with SCLC, showed a significant direct correlation between disease activity and serum LDH. In contradistinction to the present report and to at least six other reported studies23'2 5 ' 30 -32 in which elevated levels of serum LDH predict for poor outcome in patients with SCLC, Hainsworth et a133 have reported that elevated serum level at the time of diagnosis was predictive of improved survival. This finding is perplexing in that elevated
serum levels of LDH have been associated with poor prognosis in other types of neoplasms.8-12 A possible explanation, nevertheless, may be that other prognostic attributes not included in the study of Hainsworth et al may dictate prognosis in addition to LDH levels and that interactions of such variables with LDH may reflect prognosis more accurately. Hainsworth et al, in fact, state that it is conceivable that such prognostic factors in their patient population were possibly overlooked or were not possible to analyze due to the small number of patients in the study. In this regard, recent advances in the analysis of prognostic factors have focused on the use of multivariate models, which endeavor to demonstrate that several factors in addition to disease extent provide separate and additional value in the prediction of prognosis. In this respect, the correlation of serum markers with other predictors of prognosis is particularly noteworthy. Ohnoshi et a131 illustrated the value of assessing serum LDH and albumin in combination with conventional prognostic factors in a multivariate analysis. We have previously reported on the combined use of LDH and CEA as prognostic factors for SCLC.32 Both CEA and LDH levels are independent prognostic factors in SCLC; however, in comparison to LDH, which stratifies patients into three prognostic groups, the relevance of CEA was restricted to a subset of patients with LD. Interactions between LDH and other prognostic factors were used by Osterlind and Anderson, 23 Cerney et al, 24 and Albain et a125 to define prognostic risk groups by multivariate analysis. All these studies corroborated the inclusion of LDH as a valuable prognostic factor in patients with SCLC. One particular advantage of LDH is that this laboratory test is comparatively inexpensive and readily available. In this regard, Dole34 reported on a perfect correlation between the absence of bone marrow involvement and normal levels of serum LDH and recommended that the routine use of bone marrow evaluation in such patients be omitted. Similarly, in this report, only 1.7% (two of 110) of patients with normal LDH levels were found to have detectable bone marrow involvement. The corresponding rates in other series have been somewhat higher at 10% to 16%.35.36 Never-
theless, it is conceivable that in conjunction with other routinely performed laboratory parameters, the assessment of LDH may define disease burden
Downloaded from ascopubs.org by University of Newcastle Upon Tyne on April 7, 2019 from 128.240.208.034 Copyright © 2019 American Society of Clinical Oncology. All rights reserved.
960
SAGMAN ET AL
and obviate the necessity of more complex and invasive staging investigations. Consequently, the requirements for the assessment of prognosis and treatment planning could be simplified, patient discomfort minimized, and health care costs might be reduced. In conclusion, LDH is a useful independent pretreatment prognostic attribute in patients with
SCLC that correlates with extent of disease and response to treatment and survival following diagnosis. It is suggested that LDH complements other prognostic factors and, hence, may contribute to a more accurate assessment of patient status. Finally, the role of LDH in stratifying patients into treatment groups during the conduct of clinical trials should be further studied.
REFERENCES 1. Minna JD, Higgins GA, Glatstein EI: Cancer of the lung, in DeVita VT, Hellman S, Rosenberg SA (eds): Cancer, Principles and Practice of Oncology. Philadelphia, PA, Lippincott, 1985, pp 507-598 2. Morstyn G, lhde DC, Lichter AS, et al: Small cell lung cancer 1973-1983: Early progress and recent obstacles. Int J Radiat Oncol Biol Phys 10:515-539, 1984 3. Cohen MH, Ihde DC, Bunn, et al: Cyclic alternating combination chemotherapy for small cell bronchogenic carcinoma. Cancer Treat Rep 63:163-170, 1979 4. Feld R, Ginsberg RJ, Payne DG: Treatment of small cell lung cancer, in Roth JA, Ruckdeschel JC, Weisenburger TH (eds): Thoracic Oncology. Philadelphia, PA, Saunders, 1989, pp 229-262 5. Stanley KE: Prognostic factors in lung cancer, in Aisner J (ed): Lung Cancer. New York, NY, Churchill Livingstone, 1985, pp 41-66 6. Sculier JP, Feld R, Evans WK, et al: Carcinoembryonic antigen: A useful prognostic marker in small cell lung cancer. J Clin Oncol 3:1349-1354, 1985 7. Carney DN, Ihde DC, Cohen MH, et al: Serum neuron specific enolase: A marker for disease extent and response to therapy of small cell carcinoma. Lancet 1:583-585, 1982 8. Schilling RF, McKnight B, Crowley JJ: Prognostic value of serum lactate dehydrogenase level in Hodgkin's disease. J Lab Clin Med 99:382-387, 1982 9. Hagberg H, Siegbahn A: Prognostic value of serum lactic dehydrogenase in non-Hodgkin's lymphoma. Scand J Haematol 31:49-65, 1983 10. Brereton HD, Simon R, Pomeroy TC: Pretreatment lactate dehydrogenase predicting metastatic spread in Ewing's sarcoma. Ann Intern Med 83:352-354, 1975 11. Simonsson B, Grenning G, Kallander C, et al: Prognostic value of serum lactic dehydrogenase (S-LDH) in multiple myeloma. Eur J Clin Oncol 17:336-339, 1987 12. Barlogie B, Smallwood L, Smith T, et al: High serum levels of lactic dehydrogenase identify a high-grade lymphoma-like myeloma. Ann Intern Med 110:521-525, 1989 13. Feld R, Pringle JF, Evans WK, et al: Combined modality treatment of small cell carcinoma of the lung. Arch Intern Med 141:469-473, 1981 14. Clamon GH, Feld R, Evans WK, et al: The effect of adjuvant central IV hyperalimentation on the survival and response to treatment of patients with small cell lung cancer: A randomized trial. Cancer Treat Rep 69:167-177, 1985 15. Evans WK, Feld R, Murray N, et al: Superiority of alternating non-cross-resistant chemotherapy in extensive small cell lung cancer: The results of a multicentre random-
ized National Cancer Institute of Canada clinical trial. Ann Intern Med 107:451-458, 1987 16. Feld R, Evans WK, Coy P, et al: Canadian multicentre randomized trial comparing sequential and alternating administration of two non-cross-resistant chemotherapy combinations in patients with limited small cell carcinoma of the lung. J Clin Oncol 5:1401-1409, 1987 17. Kaplan EL, Meier P: Non parametric estimation from incomplete observations. Am Stat Assoc J 53:457-481, 1958 18. Peto R, Pike MC, Armitage P, et al: Design and analysis of randomized trials requiring prolonged observations of each patient II. Analysis and examples. Br J Cancer 35:1-39, 1977 19. Souhami RL, Bradbury J, Geddes DM, et al: Prognostic significance of laboratory parameters measured at diagnosis in small cell carcinoma of the lung. Cancer Res 45:2878-2882, 1985 20. Cohen MH, Makuch R, Johnston-Early A, et al: Laboratory parameters as an alternative to performance status in prognostic stratification of patients with small cell lung cancer. Cancer Treat Rep 65:187-195, 1981 21. Maurer LH, Pajak TF: Prognostic factors in small cell carcinoma of the lung: A Cancer and Leukemia Group B study. Cancer Treat Rep 65:767-774, 1981 22. Wallach SR, Royston I, Taetle R, et al: Plasma calcitonin as marker of disease activity in patients with small cell carcinoma of the lung. J Clin Endocrin Metab 53:602606, 1981 23. Osterlind K, Andersen PK: Prognostic factors in small cell lung cancer: Multivariate model based on 778 patients treated with chemotherapy with or without irradiation. Cancer Res 46:4189-4194, 1986 24. Cerney T, Blair V, Anderson H, et al: Pretreatment prognostic factors and scoring sytem in 407 small cell lung cancer patients. Int J Cancer 39:146-149, 1987 25. Albain KS, Crowley JJ, Le Blanc M, et al: Determinants of improved outcome in small-cell lung cancer: An analysis of the 2,580-patient Southwest Oncology Group data base. J Clin Oncol 8:563-1574, 1990 26. lhde DC, Makuch RW, Carney DN, et al: Prognostic implications of stage of disease and sites of metastasis in patients with small cell carcinoma of the lung treated with intensive combination chemotherapy. Am Rev Respir Dis 123:500-507, 1981 27. Byhardt RW, Hartz A, Libnoch JA, et al: Prognostic influence of TNM staging and LDH levels in small cell carcinoma of the lung (SCCL). Int J Radiat Oncol Biol Phys 12:771-777, 1986
Downloaded from ascopubs.org by University of Newcastle Upon Tyne on April 7, 2019 from 128.240.208.034 Copyright © 2019 American Society of Clinical Oncology. All rights reserved.
961
PROGNOSTIC SIGNIFICANCE OF LDH IN SCLC 28. Vezzoni MA, Lucchini R, Giardini R, et al: Lactate dehydrogenase levels in cellular extracts of human malignant lymphomas. Tumori 69:279-282, 1983 29. Ganz PA, Ma PY, Wang HJ, et al: Evaluation of three biochemical markers for serially monitoring the therapy of small cell lung cancer. J Clin Oncol 5:472-479, 1987 30. Vincent MD, Ashley SE, Smith IA: Prognostic factors in small cell lung cancer: A simple prognostic index is better than conventional staging. Eur J Cancer Clin Oncol 23:1589-1599, 1987 31. Ohnoshi T, Hiraki S, Nakata Y, et al: Pretreatment serum albumin concentration and lactic dehydrogenase activity as prognostic factors in patients with small cell lung cancer. Acta Med Okayama 36:487-490, 1982 32. Wang T, Feld R, Sagman U, et al: The combined use
of LDH and CEA as independent prognostic markers for small cell lung cancer. Proc Am Assoc Cancer Res 27:162, 1986 (abstr 641) 33. Hainsworth JD, Johnson DH, Farese RV: Limited stage small cell lung cancer: Analysis of clinical prognostic factors. South Med J 80:1518-1522, 1987 34. Dole DC: Serum lactate dehydrogenase and bone marrow involvement in small cell carcinoma of the lung. N Engl J Med 712:1262, 1985 35. Osterlind K, Hirsch FR, Pedersen AG: Serum lactate dehydrogenase and bone marrow metastasis in small cell carcinoma of the lung. N Engl J Med 313:759, 1985 36. Johnson DH, Greco AF: Serum lactate dehydrogenase and bone marrow metastasis in small cell carcinoma of the lung. N Engl J Med 313:759, 1985
Downloaded from ascopubs.org by University of Newcastle Upon Tyne on April 7, 2019 from 128.240.208.034 Copyright © 2019 American Society of Clinical Oncology. All rights reserved.
