Adapting the Lung Cancer Symptom Scale (LCSS) to mesothelioma
Using the LCSS-meso conceptual model for validation
Fax: (434) 924-2787
Abstract
BACKGROUND
The underpinning conceptual model for the Lung Cancer Symptom Scale (LCSS), an instrument used to assess health-related quality of life in patients with lung cancer, has been described elsewhere. The patient-rated scale of the LCSS was modified slightly for patients with mesothelioma (LCSS-Meso), because no other mesothelioma-specific instrument was available.
METHODS
In the current methodologic study, the authors tested the conceptual model for the LCSS-Meso. Chemotherapy-naive patients with unresectable malignant pleural mesothelioma who were participating in two clinical trials of pemetrexed (ALIMTA; Eli Lilly, Indianapolis, IN) completed the scale twice before the start of therapy and once weekly during the trials. Three time points were analyzed: baseline, Day 40, and Day 82. Poisson regression was used to determine the contribution of predictive factors (i.e., symptoms) to the summary items (symptom distress, activity level, and global quality of life).
RESULTS
The model was tested in 495 patients who had malignant pleural mesothelioma. More than 85% of patients reported pain, dyspnea, fatigue, and appetite loss. Pain, dyspnea, and fatigue were significant and stable predictors for all summary items; however, pain had a significant effect on global quality of life only through Day 40. Appetite loss was a significant and stable predictor of activity level and global quality of life. The explained variance for the model was 39–55%.
CONCLUSIONS
Further support for the content validity of the LCSS-Meso was obtained, as nearly all patients validated that the symptoms described in the scale captured their disease experience. The only exception was the hemoptysis item, which was removed based on the current large normative data set. Support for the construct validity of the LCSS-Meso also was obtained. For both mesothelioma and lung cancer, the majority of factors within the LCSS model are relevant and have the expected amount of variability. These findings support the use of the LCSS as a sensitive instrument for serial measurement during clinical trials involving patients with lung malignancies. Cancer 2004. © 2004 American Cancer Society.
Malignant mesothelioma is a highly symptomatic and rapidly progressive malignancy. Its causative relationship to asbestos exposure has made it one of the best studied malignancies from an epidemiologic standpoint. However, only recently have large, randomized, therapeutic trials become more frequent, and these have been mostly Phase II trials.1 Due to the prominent symptomatic nature of mesothelioma, it is clear that accurate evaluation of the impact of this disease and its treatment on health-related quality of life (hereafter, quality of life) and on palliative outcomes is an important goal.
The value of quality-of-life assessment using validated instruments is recognized increasingly in clinical trials. The U.S. Food and Drug Administration (FDA) and the American Society of Clinical Oncology (ASCO) Outcomes Research Committee have stated that quality of life is one of the three key endpoints for clinical trials, in addition to tumor response and survival.2, 3 Several studies have demonstrated that quality of life is a significant prognostic factor for survival in patients with thoracic malignancies; this finding has been documented best in patients with nonsmall cell lung cancer.4-6 Recent evaluations have indicated that quality of life also is an important factor in the evaluation of patients with the thoracic malignancy mesothelioma.
With no established instrument for the measurement of quality of life in patients with mesothelioma, the Lung Cancer Symptom Scale (LCSS) was identified as a potentially appropriate instrument for evaluation.7, 8 Developing an entirely new quality-of-life instrument specifically for patients with mesothelioma was not considered necessary, because this malignancy was markedly similar to lung cancer in terms of published symptom assessments and survival data. Hence, the original model underpinning the psychometric testing of the LCSS was modified slightly for testing in patients with mesothelioma.9
The study objectives were threefold: 1) to test an adapted conceptual model (the LCSS-Mesothelioma [LCSS-Meso] model), including the stability of factors during chemotherapy intervention (to provide support for construct validity), in patients with mesothelioma; 2) to identify the major presenting symptoms at baseline for a large sample of patients with mesothelioma (to provide additional support for content validity); and 3) to compare the outcomes in patients with mesothelioma in the current study with the results from an earlier study of patients with nonsmall cell lung cancer.
Once diagnosed, mesothelioma is a highly symptomatic and aggressive malignancy that has remained difficult to treat.10-13 The median survival for patients with mesothelioma is 6–9 months.1 Eighty percent of all mesotheliomas begin in the pleural space.14 Studies indicate that 70–90% of patients with mesothelioma will report pain, dyspnea, or both.15-17 Other common symptoms include fatigue and appetite loss,11, 13, 15, 18-20 with cough being reported as a symptom less frequently. All agree that it is imperative that treatment should produce symptomatic improvement and should enhance quality of life.
Decades ago, studies documented a history of exposure to asbestos in as many as 77–88% of patients with mesothelioma.20, 21 Currently, no history of asbestos exposure is noted in 30–50% of patients with mesothelioma.11 In a recent trial in France, 53% of 70 patients reported exposure to asbestos.17 The median age of patients with this disease is 60 years, and men outnumber women 5:1.11 The latency period is estimated to be 3–4 decades from asbestos exposure to the development of malignant mesothelioma.11 It is not surprising that the incidence of this disease is expected to rise over the next several decades in many industrialized countries, until legislation begins to have an effect on incidence rates.10, 11
The LCSS conceptual model used as the basis for testing the original LCSS was used again but was modified slightly as the underpinning model for the mesothelioma instrument; it was expected that after testing, the hemoptysis item would be deleted.9, 22, 23 In Figure 1, the large circle on the left operationalizes the physical dimension of quality of life for patients with mesothelioma. This dimension is captured in depth by the five major symptoms (appetite loss, fatigue, cough, dyspnea, and pain). The two boxes on the bottom row of the diagram list the less relevant dimensions that are captured by summary items (global activity status and global quality of life). These dimensions are considered less relevant, because they are less likely to be affected by a new agent. Instead of many questions related to work and home, one question is used to sum up all of the patient's normal activities, thus capturing physical, cognitive, and social (or role) function. This method of capturing less relevant dimensions in a summary item is important in the current population for two reasons: brevity and weighting. For example, a patient may consider his or her quality of life to be high because of his or her daughter's wedding, but this assessment needs to be captured in less depth, because it is less relevant to the evaluation of a new agent. These less relevant dimensions must be captured, but they should be captured using an approach that involves fewer items and produces less patient burden when the decline is rapid and when repeated measures are required for a clinical trial or for clinical practice.
In testing the original model, the three summary items were used as the dependent variables.9 These items were symptom distress (representing the physical dimension), interference with normal activity level (representing the functional dimension), and global quality of life. The individual items were used as the predictors, and a summary item was added as a predictor as the model was built from left to right. Using Poisson regression as described by McCullagh and Nelder,24 the LCSS model explained nearly half of the variance in quality of life experienced by 144 patients with nonsmall cell lung cancer during chemotherapy. The explained variance in quality of life ranged from 35% to 53% over 3 time points—baseline, Day 29, and Day 71—during chemotherapy.
MATERIALS AND METHODS
Design/Setting
The current methodologic study, which sought to obtain support for the validity of the LCSS-Meso instrument, had a longitudinal design and replicated the original study.9 For testing the modified version, the LCSS-Meso model for patients with mesothelioma, it was hypothesized that all factors would remain the same as in the original model, except for hemoptysis. Again, 3 time points were used; these time points were selected using the same rationale as in the original study but were slightly different from the original time points, due to differences in the intervention schedule (specifically, chemotherapy was administered every 21 days to patients with mesothelioma). The baseline visit was chosen to represent the period after diagnosis but before treatment. Day 19 of Cycle 2 (i.e., approximately Day 40) was chosen because it would not reveal acute toxicity but would indicate initial efficacy and sustained toxicity. Day 19 of Cycle 4 (i.e., approximately Day 82) was chosen because it also would not reveal acute toxicity but would highlight sustained efficacy and cumulative toxicity. Because of cycle delays, treatment dates were approximate, with a ±3-day window rule instituted in the current study. A modified LCSS was completed by patients twice before the start of therapy and then once weekly during the study.
The current study was associated with two prospective clinical trials of a new treatment for patients with mesothelioma. Chemotherapy-naive patients with unresectable malignant pleural mesothelioma participated in a randomized, single-blind Phase III trial comparing pemetrexed (ALIMTA; Eli Lilly, Indianapolis, IN) plus cisplatin with cisplatin or in a single-arm Phase II trial of pemetrexed monotherapy.1, 25 Pemetrexed is a novel, multitargeted antifolate with inhibitory activity against multiple enzymes; this feature differentiates pemetrexed from 5-fluorouracil and methotrexate. Pemetrexed was administered every 21 days as a 10-minute infusion, and patients received low-dose dexamethasone (4 mg orally twice daily) for 3 days, beginning 1 day before chemotherapy. Nonsteroidal antiinflammatory drugs, which commonly are used in the treatment of mesothelioma, were interrupted for 5 days, beginning 2 days before chemotherapy, due to a potential interaction (resulting in decreased renal clearance) with pemetrexed. These adjunct interventions, which may have had a temporary impact on quality of life, explain why quality of life was assessed on Day 19. In the Phase III trial, pemetrexed plus cisplatin was associated with a statistically significant survival advantage (median, 12.1 months vs. 9.3 months) as well as a statistically superior time to disease progression (median, 5.7 months vs. 3.9 months) and a statistically superior tumor response rate (41% vs. 17%). In addition, all parameters assessed with the LCSS-Meso were either superior or similar in the pemetrexed plus cisplatin group compared with the cisplatin monotherapy group.26 The Phase II trial revealed modest efficacy, with a tumor response rate of 14% and an improvement in terms of patient symptoms.27, 28 Both trials demonstrated manageable toxicity for patients who received vitamin supplementation in addition to pemetrexed.1 Both trials were multinational and thus required translation of the instrument via the standard forward/backward process as well as pilot testing. The modified LCSS-Meso was translated into the languages necessary for use at the 96 investigative sites, which were located on a total of 5 continents. After the translation process, the basic psychometric properties of feasibility, reliability, and validity were assessed, with the assessment yielding good results for this new version of the LCSS.29
Content validity involves validation of instrument content by a panel of experts in the particular construct of interest or by a patient panel.30 Construct validity addresses how well the measure captures the construct or theoretical conceptualization.30 One of the four common methods used to obtain support for construct validity is hypotheses testing of important clinical relationships drawn from the theory.31 Three criteria are needed to estimate validity: 1) multiple procedures, 2) sequential use of these procedures, and 3) assessment of validity at various stages of instrument development.32 Validity properties are a matter of degree, with no absolute standard for the magnitude of a validity coefficient; multiple testing leads to the accrual of evidence in support of the validity of an instrument.31
Patients
The study involved a multinational cohort of 495 patients who had malignant pleural mesothelioma from a total of 19 countries. Eligible patients included those with malignant pleural mesothelioma who had surgically unresectable disease, no prior chemotherapy, and a performance status ≥ 70% using the physician-rated Karnofsky performance status (KPS) scale.33
For a patient to be included in the testing of the model, none of the nine items on that patient's LCSS could be missing. Thus, of the original 512 qualified and treated patients, 17 (3%) were excluded due to missing data at baseline (9 patients from the Phase II trial and 8 patients from the Phase III trial). The overall completion rate was 90% for both studies. This rate was calculated based on the two baseline assessments plus weekly assessments until discontinuation of therapy. This combined cohort clearly met the guideline of 85% evaluability for clinical trials, as was proposed by Simon and Wittes of the National Cancer Institute Cancer Therapy Evaluation Program.34
Baseline patient characteristics are presented for the 495 patients in the current study (Table 1). Consistent with reported epidemiologic findings, there were more men (81%) than women (19%) in the cohort. At baseline, the performance status was either 80% or 90% for the majority of patients, and 79% had Stage III or IV disease. The median patient age was 61 years (range, 19–85 years). Only 13% of the 495 patients were age < 50 years, and 30% were ages 50–59 years. Almost all patients (92%) were Caucasian. Patients from the United States (24%) and patients from western Europe (49%) represented the largest subgroups.
Characteristic | Percentage of patients |
---|---|
Gender | |
Male | 81 |
Female | 19 |
KPS score | |
100 | 12 |
90 | 43 |
80 | 29 |
70 | 16 |
Disease stage | |
I | 7 |
II | 14 |
III | 33 |
IV | 46 |
Age (yrs) | |
Median | 61 |
Range | 19–85 |
Race | |
Caucasian | 92 |
Other | 8 |
- KPS: Karnofsky performance status.
Instruments
The LCSS is a site-specific quality-of-life measure that was developed specifically for patients participating in clinical trials, but it also can be used in clinical practice.7, 22 The LCSS has two brief scales: a scale for patients and an optional scale for the health professional to provide context (such as the type of pain medication used) as an observer. The LCSS was modified slightly for patients with mesothelioma by substituting lung illness for lung cancer in the instructions and in Item 7, as mesothelioma is discussed differently from the more common lung cancers. In the current study, the original nine-item patient scale was used and included six items measuring major symptoms of lung malignancies (appetite loss, fatigue, cough, dyspnea, hemoptysis, and pain) and three summary items related to total symptom distress, activity status, and overall quality of life. The patient was asked to focus on describing the symptoms of the disease itself so that descriptions of comorbidity symptoms would be limited. The time frame of the LCSS is the past day (operationalized as within the last the 24 hours), based on the rationale that scale items represent states or conditions that may fluctuate rapidly, rather than enduring traits. A straightforward example question related to the weather is provided to ensure that the patient understands how to mark a visual analogue scale (VAS) before proceeding. The interval-level VAS used contained 100 mm lines for measuring patient responses. Each item on a separate card was assigned an individual score equal to the length of the line marked by the patient, with 0 corresponding to the best rating and 100 representing the worst rating. The average of the aggregate score on all nine items was reported as the total score. In addition, a subscore representing the mean of all six major symptom scores (the average symptom burden index), the single global quality-of-life item, and/or individual items can be used to assess specific areas of change. The initial mean administration time has been reported as 8 minutes (which included the time required to provide instructions) and generally decreases to 3–5 minutes with repeated administrations. The readability index for the patient scale is a second-grade reading level.
Using the standard forward/backward translation process, there are currently 42 translations of the original instrument that have been tested in patients with lung cancer. Previously reported α coefficients for the translations used in the study ranged from 0.80 to 0.89.29 The original LCSS has been tested in > 1000 patients with lung cancer. The psychometric properties of feasibility, reliability (coefficient α for internal consistency and test-retest for stability), and validity (contrasted groups approach, hypotheses testing, multitrait-multimethod approach, and criterion-related validity) have been well documented.7-9, 22, 23 Normative data also have been published for the original measure.35
With regard to reliability, the Cronbach α coefficient was 0.85 for the total scale in a subgroup of 435 patients with mesothelioma in the current study. As another measure of reliability, the stability (or reproducibility) of the modified scale also was examined using the test-retest method. Using a testing interval of 3–5 days before the start of therapy, the Pearson correlation coefficient was found to be 0.86 for a subgroup of 376 patients in the current study. Using the guideline put forth by Nunnally and Bernstein that a reliability coefficient ≥ 0.70 is acceptable for new measures,36 the modified LCSS was judged to be reliable for this group of patients with mesothelioma.
The observer-rated Karnofsky scale is a well accepted and widely used instrument in oncologic practice.32 It is used to measure physical functioning based on level of activity, symptoms of disease, and amount of assistance required. The Karnofsky scale categorizes a patient's activity level using an 11-point ordinal level scale, with scores ranging from 0% (dead) to 100% (normal activities with no signs or symptoms of disease).
Procedures
Approval from the human investigations committee was obtained at each investigative site before the start of the study, and written informed consent was obtained from each participant before the baseline visit. Site personnel were instructed regarding proper administration of the LCSS before the enrollment of any patients. Administration of the LCSS was planned to coincide with other protocol procedures whenever possible and to precede any procedures or discussions that may have influenced results. Site personnel administered the LCSS, measured the VAS scores, and recorded responses on a case report form.
Analysis Method
Three analyses were completed to determine the contribution of the predictive factors in the LCSS-Meso model for quality of life. The dependent variables in the model were the three summary items: symptom distress, interference with activity level, and global quality of life. The individual symptom items were used as predictors. Using the criterion put forth by Schroeder in 1990 that correlations > 0.85 were an indicator of multicollinearity,37 the process of adding the dependent variables (summary items) as predictors as the model builds from left ot right was not included for this analysis (as in the original model testing). In addition, as in the testing of the original model, Poisson regression was used to analyze the data at the three time points.24 Poisson regression is a parametric approach based on a mathematical model. It is used when the assumption of normality is not met and assumes that the data are distributed in a Poisson distribution, in which data are treated as counts. Parameter estimates and corresponding standard errors were estimated by maximum-likelihood methodology. The Wald chi-square statistic was used to assess the relative importance (i.e., predictive power) of the regression variables. Wald chi-square statistics (each with 1 degree of freedom) were calculated as follows: (value of the regression coefficient) ÷ (estimated standard error of the coefficient)2. Thus, the larger a coefficient was relative to its standard error, the larger its chi-square statistic was, corresponding to a greater relative importance for the regression variable in question. All tests of significance (P < 0.05) were two-tailed.
RESULTS
The major presenting symptoms at baseline are illustrated in Figure 2. Underscoring the fact that mesothelioma is a highly symptomatic disease are the findings that at least 85% of the 495 patients experienced each of the 4 most common symptoms: pain, dyspnea, fatigue, and appetite loss. Of the 495 patients with mesothelioma, 92% reported ≥ 3 symptoms (Fig. 3), compared with 85% of the previous study cohort of 144 patients with nonsmall cell lung cancer.
Because there is little or no published information regarding the severity of symptoms for patients with mesothelioma, distributions for each symptom score were obtained using box-and-whisker plots for both the mesothelioma cohort and the nonsmall cell lung cancer cohort (Fig. 4). The horizontal line in Figure 4 denotes the median, the edges of the box represent the 25th and 75th percentiles, and the whiskers provide the range of scores. These data demonstrate the symptomatic nature of mesothelioma and how the current study cohort differs from the cohort of patients with nonsmall cell lung cancer, who had slightly lower performance status scores. Among the 144 patients with nonsmall cell lung cancer, the baseline KPS score was 100% in 3 patients (2%), 90% in 40 patients (28%), 80% in 38 patients (26%), 70% in 47 patients (33%), and 60% in 16 patients (11%). Overall, pain was more common among patients with mesothelioma, but cough, fatigue, and dyspnea were less common; the incidence of appetite loss was roughly the same in both cohorts.
Symptom distress results for patients with mesothelioma followed the expected trend, with the exception of hemoptysis (Table 2). At baseline, there were four significant predictors: pain, dyspnea, fatigue, and hemoptysis. When three outliers were removed, hemoptysis was no longer statistically significant (P = 0.41). Three of 4 factors also were significant at Day 40 and at Day 82, with the order of the magnitudes of the contributions being consistent with the order observed at baseline. The explained variance (R2) increased going from one time point to the next (43% at baseline, 52% at Day 40, and 57% at Day 82).
Symptom | P valuea | ||
---|---|---|---|
Baseline (n = 495) | Day 40 (n = 363) | Day 82 (n = 271) | |
Pain | <0.001 | <0.001 | <0.001 |
Dyspnea | <0.001 | <0.001 | <0.001 |
Fatigue | <0.001 | 0.005 | 0.003 |
Hemoptysis | 0.005 | — | — |
Explained variance (R2) | 0.43 | 0.52 | 0.57 |
- a Values are presented in order of magnitude (chi-square test; Wald statistic) at baseline. The order was consistent (relative to baseline) at Days 40 and 82.
For patients with mesothelioma, the effects associated with the second component of the model, interference with activity, also were as expected (Table 3). There were four significant predictors at baseline: fatigue, dyspnea, pain, and appetite loss. All four predictors also were significant at Day 40 and at Day 82, but only at Day 40 was the order consistent with the order observed at baseline. By Day 82, the predictive power of pain had decreased, possibly due to decreased pain as a result of tumor response. (Mean scores for pain decreased relative to baseline, whereas other predictor scores remained stable or increased.) Similarly, the effects of appetite loss may be relatively minor, due to emesis control by the physician. Again, the explained variance increased across the three time points examined (37% at baseline, 55% at Day 40, and 60% at Day 82).
Symptom | P valuea | ||
---|---|---|---|
Baseline (n = 495) | Day 40 (n = 363) | Day 82 (n = 271) | |
Fatigue | <0.001 | <0.001 | <0.001 |
Dyspnea | <0.001 | <0.001 | <0.001 |
Pain | <0.001 | <0.001 | 0.049 |
Appetite loss | 0.015 | 0.003 | 0.046 |
Explained variance (R2) | 0.37 | 0.55 | 0.60 |
- a Values are presented in order of magnitude (chi-square test; Wald statistic) at baseline. The order was consistent (relative to baseline) at Day 40.
The factors contributing to global quality of life for patients with mesothelioma also were as expected (Table 4). Again, there were four significant predictors at baseline: fatigue, dyspnea, pain, and appetite loss. All four predictors also were significant at Day 40, but the order of contribution was not consistent with the order observed at baseline. The observed significance of appetite loss (as evidenced by the increased mean scores for this parameter) may have been attributable to the nausea and emesis associated with cisplatin. Only three factors were significant and stable predictors by Day 80, with pain losing significance (as evidenced by decreasing mean pain scores), most likely due to pain relief associated with tumor response. As in the other two parts of the model, the explained variance increased across the three time points (39% at baseline, 52% at Day 40, and 55% at Day 82).
Symptom | P valuea | ||
---|---|---|---|
Baseline (n = 495) | Day 40 (n = 363) | Day 82 (n = 271) | |
Fatigue | <0.001 | <0.001 | <0.001 |
Dyspnea | <0.001 | <0.001 | 0.001 |
Pain | <0.001 | <0.001 | — |
Appetite loss | 0.003 | <0.001 | 0.010 |
Explained variance (R2) | 0.39 | 0.52 | 0.55 |
- a Values are presented in order of magnitude (chi-square test; Wald statistic) at baseline.
DISCUSSION
Beyond documentation of symptoms identified in the literature, the current findings provide further support for the content validity of the modified quality-of-life measure, the LCSS-Meso, for patients with mesothelioma. Nearly all of the 495 patients with mesothelioma validated that the symptoms included in the scale captured their disease experience. The only exception was the hemoptysis item, which subsequently was removed from the instrument based on this large normative data set.
Support for the construct validity of the instrument also was obtained through testing of the underpinning model in patients with mesothelioma. The relationships of the predictors to the dependent variables were as expected. Pain, dyspnea, and fatigue were significant and stable predictors of symptom distress for patients with mesothelioma at all three time points (baseline, Day 40, and Day 82). Fatigue, dyspnea, pain, and appetite loss were significant and stable predictors of activity level at all three time points. Fatigue, dyspnea, and appetite loss were significant and stable predictors of global quality of life at all three time points, with pain being a significant contributor only through Day 40. Patients who contributed data at the later time points were more likely to have achieved at least a minor tumor response. Thus, pain scores probably were less significant as a result of tumor response.
As expected, the explained variance with respect to global quality of life was similar in both models (lung cancer and mesothelioma) throughout chemotherapy (35–53% for 144 patients with nonsmall cell lung cancer; 39–55% for 495 patients with mesothelioma); nonetheless, there were differences between the two models. The explained variance not only was reproducible over all 3 time points for the cohort of 495 patients with mesothelioma but also increased consistently in all 3 parts of the model. In the original testing of the model in 144 patients with nonsmall cell lung cancer, this was not the case; the amount of explained variance was reproducible and fluctuated as expected but did not increase consistently. By the last time point, the LCSS-Meso model explained half of the variance (55%) in terms of quality of life for patients with mesothelioma, compared with roughly one-third (35%) of the variance for patients with nonsmall cell lung cancer. This difference probably was attributable to the slightly lower performance status at baseline in the nonsmall cell lung cancer cohort, the slight differences in the time points investigated, and the large difference in cohort size between these two methodologic studies.
Compared with previous models for patients with advanced malignancies, both the LCSS model and the LCSS-Meso model explained a reasonable amount of the variance for each patient cohort, thus providing support for the construct validity of the two measures. The results of the current study and of previous LCSS model testing are similar to those reported by Mor, who also examined the stability of factors in predicting quality of life.38 The dependent variables used in that study (disease symptoms, physical functioning, emotional functioning, and social functioning) were similar to those used in the LCSS and LCSS-Meso models. The explained variance reported in the three large cohorts spanning the full continuum of malignant disease reflected disease progression—for 397 newly diagnosed patients, the explained variance was 72%; for 194 chemotherapy recipients, the explained variance was 42%; and for > 2000 patients with terminal disease, the explained variance was only 22%. Mor concluded that the primary predictor was physical functioning, with this dimension constraining psychosocial functioning. Mor also found that patients with lung cancer in the newly diagnosed group had the lowest quality-of-life scores, because of the rapid progression of their disease.
The results of the current study indicate that lung cancer and mesothelioma are similar in terms of symptomology and that a modified LCSS model, the LCSS-Meso model, is applicable and yields reproducible results for patients with mesothelioma. In patients with mesothelioma and patients with nonsmall cell lung cancer, the majority of factors in the LCSS-Meso model and the LCSS model, respectively, were relevant and exhibited expected fluctuations; these findings support the use of the LCSS and the LCSS-Meso as sensitive instruments for serial measurement in clinical trials and practice involving patients with lung malignancies.