Improving the performance status in advanced non-small cell lung cancer patients with chemotherapy (ImPACt trial): a phase 2 study

This phase II trial is designed to test whether the performance status (PS) of metastatic non-small cell lung cancer (mNSCLC) patients (pts) can improve with chemotherapy if their poor PS (Eastern Cooperative Oncology Group (ECOG) PS of ≥ 2) is due to disease burden rather than comorbidities. Age18–65 years, Charlson’s comorbidity index < 9, serum albumin ≥ 3.5 g/dl, adequate bone marrow and organ function, & ECOG PS ≥ 2 as judged by the worst score of three independent physicians were administered 3 doses of weekly paclitaxel at 60 mg/m2/dose. The primary endpoint was an improvement in ECOG PS by 1 point at 4 weeks; others: toxicity (CTCAE v 5.0), quality of life (QoL) assessment at baseline and 4 weeks by EORTC QLQ-C30 and EORTC QLQ-LC13. Optimal Simon’s 2-stage design was used. Forty-six patients were included with a median age of 56 years (interquartile range, IQR 54–59), 12 (26%) had comorbid conditions, and 87% with ECOG PS 3/4. PS improved in 11 pts at 4 weeks and in 7 beyond this time point. Grade 3/4 toxicities are seen in 20% (most common: anemia and diarrhea). At a median follow-up of 4.8 m (95% CI 3.27–14.9), the median progression-free survival and overall survival were 3.3 months (95% CI 2.36–5.6) and 6.8 months (95% CI 2.47–8.8), respectively. QoL improved for global QoL, role functioning, pain, dyspnea, insomnia, pain in the chest, pain in other parts, and worsened for alopecia and sore mouth. Abbreviated chemotherapy is a useful, well-tolerated strategy in carefully selected poor PS mNSCLC patients that can improve PS and QoL. Clinical trial information: CTRI/2020/01/022617.


Introduction
The management of metastatic non-small cell lung cancer (NSCLC) has undergone a complete transformation in the era of precision oncology. About 60% of patients of metastatic NSCLC (mNSCLC) harbor a driver mutation (Kris et al. 2014;Suidan et al. 2019), the majority of which have available targeted drugs that have improved outcomes as compared to standard chemotherapy (Tan and Tan 2022). Unprecedented 5-year overall survival (OS) of up to 30% have been recently seen with immunotherapy drugs, with and without chemotherapy, in driver mutation negative mNSCLC (Reck et al. 2021).
These remarkable advances, however, fail to ameliorate the prognosis of patients with poor performance (PS) status i.e., those with an Eastern Cooperative Oncology Group (ECOG) (Oken et al. 1982) of 2 or more. Poor PS is seen in 30-40% of mNSCLC patients, which is a substantial number considering the high incidence of lung cancer overall (Radzikowska et al. 2002;Lilenbaum et al. 2008). While data supports the use of tyrosine kinase inhibitors for treating mutation-driven mNSCLC with poor PS (Inoue et al. 2009;Singh et al. 2022), there is equipoise in the management of this subgroup in driver mutation-negative cases.
Most studies have historically excluded patients with ECOG PS 2 or more (Bonomi et al. 2018). The majority of immunotherapy trials have stringent inclusion criteria and include ECOG 0-1 patients exclusively (Tang et al. 2022). Emerging data of immunotherapy in ECOG PS 2 or more in two recent studies have shown some promise (G et al. 2020, p. 2;Lee et al. 2022); a few others, however, demonstrate reduced efficacy (Fg et al. 2020, p.). The financial toxicity of expensive drugs like immune checkpoint inhibitors is tremendous. Less than 2% of total patients in whom immunotherapy was indicated are able to receive it in low to middle-income countries (Patel et al. 2022). Limited studies have previously shown that conventional chemotherapy in metastatic NSCLC improves survival irrespective of PS of the patient (Non-Small Cell Lung Cancer Collaborative Group 2010). Furthermore, palliative chemotherapy may lead to improvement in general condition, performance status, and quality of life (QoL) (Sculier et al. 2007;Kancharla et al. 2020).
PS is an overall measure of the functional status of a patient determined by comorbidities, nutritional status, and disease burden. In addition to the lack of distinctions based on etiology, the measurement of PS has been shown to have inter-observer bias (Sørensen et al. 1993;Ando et al. 2001;Blagden et al. 2003). It is possible that patients having poor PS primarily due to tumor-related symptoms would benefit more from chemotherapy as compared to other patients. The validity of excluding patients from potentially beneficial treatment based on PS alone has been questioned by experts (Gridelli et al. 2022).
We designed a single-arm phase 2 trial to assess an adaptive chemotherapy strategy of initial low-dose chemotherapy in a carefully curated patient population with poor PS with escalation to standard platinum doublet chemotherapy in those who improve.

Materials and methods
This was an investigator-initiated, single-arm phase 2 clinical trial conducted at an academic institution in India. The study was conducted in accordance with Good Clinical Practice and approved by the institute ethics committee (IECPG-63428.11.2019 dated 29.11.2019). It was registered with the Clinical Trial Registry of India (CTRI/2020/01/022617).

Patient eligibility
All patients of histologically proven stage IIIC/IV NSCLC (AJCC 8th ed) (Rami-Porta et al. 2014), aged 18-65 years, with ECOG PS of 2 or more, were eligible for this study. Inclusion criteria also encompassed adequate bone marrow, liver, and kidney function: defined as hemoglobin of ≥ 8 g/ dl, platelet count of ≥ 1,00,000/mm 3 , ANC of ≥ 1500cells/ mm 3 , bilirubin less than 2 times ULN, transaminases less than 3 times ULN, and an estimated GFR of ≥ 30 ml/min/ BSA, respectively. Charlson's comorbidity index (CCI) score (Charlson et al. 1987;Zhao et al. 2017) was calculated at baseline. Participants should have had a CCI score of less than 9 with any comorbid conditions under control and a serum albumin of ≥ 3.5 g/dl.
Patients who were pregnant or lactating, had secondary malignancy, had HIV or hepatitis B or C positive status, known driver mutation-positive status for EGFR/ALK/ ROS1, and symptomatic untreated brain metastasis was excluded from the study. Unknown/pending driver mutation status patients were allowed to be enrolled, and if they were found to have these alterations during the course, they were excluded from the study. Treated brain metastasis or those with asymptomatic brain metastasis were allowed. Informed consent was obtained from all participants.

Treatment/intervention
Enrolled patients received paclitaxel at 60 mg/m 2 of body surface area (BSA) intravenously (IV) over 1 h every seven days (± 1 day for holiday/Sunday) for three doses. Pre-and post-medication were as per institute protocol. In addition, palliative radiotherapy and supportive treatment for symptom palliation were provided at the treating physician's discretion.

PS assessment
ECOG PS assessment was done by three physicians other than the primary investigator: a consultant medical oncologist, one medical oncology fellow, and one palliative medicine physician. The assessment was performed independently from one another. The worst of the three scores was considered. This was done to mitigate bias in the evaluation of performance scores. ECOG PS was recorded at baseline and 4 weeks from the start of therapy.

Evaluation
Comprehensive history and examination, complete blood count (CBC), liver and kidney function tests (LFT and KFT), serum LDH, peripheral smear, and staging by contrastenhanced computed tomography (CECT) of chest, abdomen, and pelvis or Whole body PETCT were recorded at baseline. Before each dose of chemotherapy, patients were examined physically, and CBC/LFT/KFT for their general condition and toxicities were recorded with their grade of severity.

Quality of life assessment
QoL assessment was performed using two European Organization for Research and Treatment of Cancer (EORTC) questionnaires: a core questionnaire (QLQ-C 30) and a Lung cancer focussed questionnaire, LC-13 (2017). The QLQ C-30 thirty-item cancer-specific questionnaire consists of five functional domain scales (physical, role, emotional, social, and cognitive), global QoL, and symptom scales. LC-13 has questions on lung cancer-associated symptoms and adverse effects related to treatment QoL was measured at baseline and at 4 weeks for all patients.
Permission for the use of both questionnaires was obtained for Hindi and English (EORTC request ID: 63,617). Interpretation was as per the EORTC manual. The patient completed the questionnaires at baseline and four weeks.

Management of toxicity
Toxicities were measured using the CTCAE v5.0 scale (2017). All grade 1 toxicity was managed conservatively. For grade 2 toxicities, we provided symptomatic management until they were reduced to grade 1 or better. For grade 3/4 toxicities, patients were not administered further chemotherapy and were provided the best supportive care out of protocol.

Study endpoints
The study's primary endpoint was an improvement in PS by 1 point at the end of 4 weeks after receiving three doses of weekly paclitaxel. Secondary endpoints included assessment of adverse events, QoL, progression-free survival (PFS), and overall survival (OS).
PFS was defined as the time from the first day of treatment until progressive disease/death from any cause, whichever came first. Surviving patients who did not progress were censored for PFS as of the date of their last assessment or (date of data lock, 31st December 2021), whichever was earlier. OS was defined as the time from the first day of treatment until the date of death from any cause.

Statistical analysis
We assumed an improvement in PS by one point in 25% of patients with chemotherapy (alternative hypothesis) vs. 10% with best supportive care (BSC) alone (null hypothesis). Accrual of forty-three patients was required for a type 1 error of 0.048 and a power of 80% to detect whether PS improvement with paclitaxel would be more than 25%. We used the Simon 2-stage optimal design (Simon 1989) with an initial 18 patients enrolled and evaluated for improvement in PS. The trial would continue only if three or more patients showed an increase in ECOG PS by one point at the end of one month. The null hypothesis would have to be rejected if eight or more responses were observed.
Data entry was done using Microsoft Excel, and analysis was done with STATA 14 (StataCorp™, College Station, TX, USA). Baseline categorical variables were analyzed using the Chi-Square test/Fischer Exact test. Proportions were calculated and reported with 95% confidence intervals (CI). Non-parametric variables were analyzed using paired t-test/ Mann-Whitney test for paired data and Wilcoxon sign rank/independent t-test for unmatched data. Survival was estimated by the Kaplan-Meier's method and the log-rank test was used to identify prognostic factors. Univariate and multivariate analyses were performed to identify predictive factors for survival. Cox proportional hazards model was used to calculate the hazard ratio.

Primary endpoint: improvement in PS
The primary objective of improvement in PS by 1 point on the ECOG scale at four weeks was seen in 11 patients (24%) overall. Out of 16 patients enrolled in the first stage of optimal Simon 2 stage design, 5 showed an improved PS, resulting in the study progressing to stage 2. Further, 30 more patients were enrolled, of which 6 demonstrated an improved ECOG PS. Two of the 11 patients were subsequently treated out of the study. One was detected to have EGFR Exon 19 deletion and was started on Gefitinib. Another patient had improvement in PS, from ECOG PS of 4 to ECOG PS of 3, and was continued on single-agent paclitaxel. The remaining nine patients went on to receive paclitaxel carboplatin doublet. A stable ECOG PS was seen in 21 patients, while it deteriorated in 14 (Fig. 2).
Further, seven patients showed an improvement in ECOG PS beyond the study endpoint of 4 weeks and successfully received standard platinum doublet therapy but outside of the study.

3
Toxicity Paclitaxel was well tolerated in this study, with 80% of patients able to complete three weekly doses. A total of 12 patients (26%) had dose delays/interruptions of chemotherapy. The most common cause of chemotherapy delay was grade III diarrhea, seen in three patients. Two patients had a delay due to grade III anemia. One patient each had worsening PS, and subacute intestinal obstruction secondary to opioids needed admission. Two patients were lost to follow-up. Two had sudden death at home due to unknown causes. One patient delayed chemotherapy due to severe pain, which was optimized by the palliative team, after which he received the second dose. All grade toxicities as per CTCAE v 5.0 were seen in 85% of patients, the most common being nausea/vomiting (50%) and anemia (45.6%). Grade 3 or greater toxicities were seen in 20% of patients; anemia (13%) followed by diarrhea (8.7%) was the most frequent (Table 2).

Survival outcomes
At a median follow-up of 4.8 m (95% CI 3.27-14.9), the median PFS of the entire ITT population was 3.3 months (95% CI 2.36-5.6), with a 6-month PFS rate (defined as patients without progression/death six months from the start of therapy) of 26% (95% CI 11-44). The median overall survival (OS) was 6.8 months (95% CI 2.47-8.8) as of 31st December 2021 ( Supplementary Fig. B1). Nine patients received doublet due to PS improvement at four weeks. An additional seven patients received doublet out of the study due to PS improvement at a later time point. At the time of publication, with a date cut off of 31st Aug 2022, seven patients are alive, and four have been lost to follow-up. The 6-month progression-free survival rate of patients receiving doublet is 31.25% (95% CI 11.39-53.65), and the median OS is 11.1 months (7.93-not reached).

Quality of life
QoL was measured using EORTC QLQ-C30 and LC-13 questionnaires. The 4th-week QoL measures were better for 'global QoL,' 'role functioning,' 'pain,' 'dyspnea,' and 'insomnia' on the EORTC QLQ-C30 and scores for 'chest pain' and 'pain elsewhere' on the LC-13 survey. An increase in 'sore mouth' and 'alopecia' was also noted. (Supplementary Tables A1 and A2) Violin plots of the QoL variables demonstrate that in addition to improvement of the median of these scores, the population density distribution shifted in a positive direction, with a greater density of people (wider width of the graph) at higher scores for functional scales and lower scores for symptom scales, respectively (Figs. 3 and 4).
In patients whose ECOG PS improved, the reduction in symptom burden of 'nausea/vomiting', 'cough', and 'dysphagia' was more significant than those without ECOG PS improvement. In addition, there was a trend to greater benefit in 'global health status' and 'pain in arm or shoulder' and 'pain in the chest' in those for whom ECOG PS improved (Supplementary Tables A3 and A4; supplementary Fig. B2).

Post hoc analysis
The intraclass correlation of performance status assessment among the three assessor categories, medical oncology fellow, consultant medical oncology, and palliative care physician, was calculated. There was moderate to good reliability at both baseline (Cronbach's α: 0.82, 95% CI 0.69-0.89) and four weeks (Cronbach's α: 0.9, 95% CI 0.81-0.94). Eligible for phase 2 (PS improved by 1 point AND current ECOG PS < = 2) 10 Received TP at least 1 cycle 9 (20%) 1 3 Cox proportional hazards analysis was done for PFS, and OS and logistic regression analysis for the primary end point of improvement in PS. For the improvement in PS, the presence of squamous histology predicted for better response on multivariate analysis (supplementary table A5). Neutrophil-lymphocyte ratio (NLR) greater than the median predicted for worse PFS and OS, and CCI greater than 8 predicted for worse PFS on multivariate analysis (supplementary table A6 and A7). On excluding the seven patients whose PS got better at a later endpoint beyond four weeks (supplementary Fig. B3), for the remaining 39 patients, improvement in ECOG PS by one point was indicative of better OS (improved PS: 6.8 months (95% CI 1.5-NR) vs. not improved: 2.37 months (95% CI 1.1-7.67).

Discussion
This is the first study, to our knowledge, that evaluated a management plan curated to the etiology of poor performance status in NSCLC patients in the form of an adaptive design that optimizes inclusion of those patients with a poor PS secondary to disease burden. We demonstrated that the PS improved in 24% of patients at the pre-decided endpoint and seven more patients at a later endpoint, with acceptable toxicity and better QoL. The concept of using a 'pre-phase' therapy in patients with the poor general condition to achieve the dual benefit of therapeutic efficacy and acceptable toxicity in patients who are not considered fit for the standard of care therapy has been explored previously in lymphomas and germ cell tumor (Lakshmaiah et al. 2018;Tryakin et al. 2018), but not for lung cancer. Overall, 39% of patients had an improvement in PS in our study, which is in line with previously available data (Leong et al. 2007;Sculier et al. 2007;Kancharla et al. 2020). This is despite a significant proportion of PS 3 and 4 patients in our study and indicates that palliative chemotherapy can positively impact carefully selected patients in this population.
The inclusion criteria used in this study reflect the selection process of excluding those with multiple comorbid conditions and poor nutrition. We included patients with a CCI score (Charlson et al. 1987) of less than 9. Metastatic NSCLC scores 6 on this scale; this limited our patients to 2 comorbidities if less than 60 years of age and one comorbidity between 60 and 65 years. We required normal to near normal organ function and albumin of ≥ 3.5 g/dl.
Paclitaxel was relatively well tolerated with no new safety signals and is analogous to previous studies in this patient population (Reynolds et al. 2009;Morabito et al. 2013;Zukin et al. 2013). Adverse events of grade 3-4 CTCAE v 5.0 was noted in 20% of patients. Anemia and diarrhea were most commonly seen. This is comparable to the available trial data for toxicity for PS 2 NSCLC patients (32,33) and even compares favorably to the side effect profile seen when this group is treated with immunotherapy (Felip et al. 2020, p. 171;G et al. 2020, p. 2).
QoL improved in the overall population in several domains (global quality of health, role functioning, pain, insomnia, and dyspnea). This came at the cost of increased mucositis and alopecia, which were expected toxicities with chemotherapy. There was a trend towards better functioning scores and reduced burden of symptoms between those who had their performance score improve at four weeks and those who did not. Coughing and dysphagia showed a statistically significant difference in the two groups, whereas pain and global health status showed a trend toward benefit. This could reflect the benefit of chemotherapy independent of improvement in performance score at four weeks, although the small sample size impedes interpretation. In our view, improving the QOL with just three doses of weekly paclitaxel in all patients, and not just those whose performance status improved, represents the vital role chemotherapy can play in poor performance status advanced NSCLC patients. Indeed, frequent visits to the hospital for weekly administration of chemotherapy did not seem to have any detrimental impact. Other studies have shown that chemotherapy improves QoL scores in metastatic NSCLC patients, with greater benefit in QoL in poor PS vs. good PS patients (Leong et al. 2007;Helbekkmo et al. 2007;Belani et al. 2008).
The inter-observer performance status assessment at both time points by three assessors had moderate to good reliability, which was at par with the existing information on interobserver correlation at the physician level on ECOG performance status (Sørensen et al. 1993;Myers et al. 2010).
This study's median PFS and OS are comparable to existing literature focusing on patients with ECOG PS > 1(Le Chevalier et al. 1994;Kosmidis et al. 2007;Reynolds et al. 2009;Zukin et al. 2013). However, in striking contrast to these studies, the current study included a predominance of ECOG 3-4 patients (87%). This indicates a benefit of chemotherapy irrespective of PS in NSCLC patients. Recent immunotherapy studies in poor PS patients have shown similar safety and less efficacy than fit patients (G et al. 2020, p. 2; R et al. 2021, p.) Surprisingly, with the caveat of cross-trial comparisons, immunotherapy does not yield better results than chemotherapy in this population (Facchinetti et al. 2020;Sehgal et al. 2021).
We could not find any predictive factors that could predict PS improvement beyond squamous histology, which had higher odds than others. The timing of the assessment may need adjustment, as some patients showed improvement after the cut-off of 4 weeks. Possibly, evaluation at six weeks would have been more appropriate. Indeed, some retrospective studies have shown more remarkable improvement in poor PS patients after 3-4 cycles of chemotherapy (Leong et al. 2007;Sculier et al. 2007;Kancharla et al. 2020).

Conclusions
To conclude, the use of palliative chemotherapy, such as weekly paclitaxel, can improve the performance status and QoL in advanced NSCLC patients with poor PS with acceptable toxicity and survival. The role of such improvement as a bridge to standard platinum doublet chemotherapy is an attractive option in this setting and further evaluation is needed in a randomized controlled trial setting.