Real-World Testing Practices, Treatment Patterns and Clinical Outcomes in Patients from Central Eastern Europe with EGFR-Mutated Advanced Non-Small Cell Lung Cancer: A Retrospective Chart Review Study (REFLECT)

Reviewer 1

This is a pooled retrospective chart review of patients diagnosed with NSCLC to assess and describe treatment practices in TKI therapies. The paper is very clearly organized and well written; the tables and figures are appropriate and informative.

The primary objectives as stated were ..."the proportion of 1G/2G EGFR TKIs used as 1L therapy in patients with EGFRm NSCLC, real-world progression free survival (rwPFS) on 1L and treatment strategies in patients receiving 2L therapy." Figures 1 and 2 does not make sense with the stated objective. If the interest was to assess real world progression free survival then a Kaplan Meier plot would have been the most informative and in line with the objectives. I don't really understand the listing of the proportion blocks for describing this type of survival in any real informative way. Additionally, since multiple combinations of therapy were used it's difficult to discern how different therapies contribute to survival versus treatment changes. The authors should work with a statistician to design an analysis where this can be assessed or rewrite the objectives to just state that they are interested in describing treatment modalities by patient and tumor characteristics. The listed outcomes also make no sense with what is listed as the objective...as it is written the exposures are the combinations of the treatment modalities 1L, 2L (and 3L?) and the outcomes are death, progression, and treatment modality change. Although the figures are well done, it's difficult to discern any take home message from the data as presented. Please either exclude or separate patients that are on multiple treatments at baseline or censor them during follow-up when treatment modality changes. I have significant concerns about patient and treatment heterogeneity. Since the majority of patients are diagnosed at stage IV (83%) it makes sense to separate and focus on these or at the minimum do a sensitivity analysis where the cohort is separated into binary groupings: metastatic and non-metastatic. As 94% of tumors were adenocarcinomas, it also makes sense to do a focused analysis on this group since pathology and treatment can vary widely by histology. Was there additional testing done for mutation during follow-up? I would expect treatment guidelines to vary by clinic/country...please state whether this was considered or not.

Thank you very much for your detailed feedback and insightful recommendations.

Indeed, the REFLECT study aimed to provide as many insights as possible from the real-world practice, and it included multiple objectives. However, the main focus of the study was on patient attrition rates following first-line first- or second-generation (1L 1G/2G) EGFR TKIs. The study aimed to understand mainly the real-world sequencing, EGFR T790M testing rate and how the 1G/2G EGFR TKI therapy were being used in small and medium-sized countries from Europe, and to bring thus information from various geographies. In a region where registries are not uniformly implemented and/or data are collected in a less standardized manner, this is the first analysis of the population of patients with confirmed EGFR-mutated non-small cell lung cancer (NSCLC) receiving 1L 1G/2G EGFR TKIs in several Central Eastern European (CEE) countries, in a period when this was the standard treatment. Indeed, in this secondary manuscript we focused on treatments distribution and attrition rates across lines to build evidence on such outcomes for further referencing in this region, especially that the standard of care has changed, and data is more scarce compared to other regions or Western European countries. In addition, as you already remarked, differences across countries exist and due to this fact, we did not find relevant the Kaplan-Meier curve for the pooled data of CEE patients, especially when the primary manuscript of the study provided details by country (reference 24, Addeo A et al, Ther Adv Med Oncol. 2021). The objectives and outcomes have been defined at main study level, and the analyses have been conducted in a centralized way, which is currently limiting our possibility to address all suggestions in a timely manner fashion. However, building on your great feedback, we believe it may be more relevant to include additional details on rwPFS data by CEE country in the Results section and include a note on this in Discussions, to be thus more informative (Note: for consistency in reporting, we also described the overall survival data by country). We have included in italic font in the column describing the changes the new and/or revisited text (lines 211-219 and lines 336-339).

We are extremely thankful for the recommendations made and we will take them into account for further analyses in this cohort. For the time being, these are limitations; therefore, we have revisited this section and expanded on these topics. We have included in italic font in the column describing the changes the new and/or revisited text (lines 371-372).

The study collected mainly information on the T790M testing results; we gathered information on other mutations as well during initial EGFR testing and follow-up, but data were limited (ie, PD-L1 expression positive for <3% of patients, 1 patient with ALK rearrangement and 1 patient with TP53 variant mutations reported, whereas at the time of T790M testing, <1% patients were identified with PD-L1 positive expression and 1 patient with MET amplification).

Lines 211-219 (Results)

Median rwPFS on 1L EGFR TKI across the pooled CEE cohort was 14.0 (95% CI: 12.6; 15.6) months, with the following results at country level: Bulgaria 16.3 (95% CI: 11.5; 22.6) months, Poland 12.9 (95% CI: 10.3; 15.4) months, Romania 12.0 (95% CI: 10.3; 15.6) months and Slovenia 15.6 (95% CI: 12.6; 19.2) months, respectively.

At the end of data collection period, 225 (58%) deaths have been recorded. Median OS from the start of 1L EGFR TKI therapy was 26.6 (95% CI: 24.1; 29.0) months across the entire CEE cohort, with the following results by country: Bulgaria 23.4 (95% CI: 18.2;20.2) months, Poland 26.2 (95% CI: 18.0; 29.7) months, Romania 26.4 (95% CI 22.4; 34.2) months and Slovenia 28.9 (95% CI: 25.0; 34.3) months, respectively.

Lines 336 – 339 (Discussion)

However, rwPFS at country level in this CEE cohort varied, ranging from 12.0 months in Romania and 12.9 months in Poland to 15.6 months in Slovenia and 16.3 months in Bulgaria, differences probably explained by patients’ heterogeneity and local characteristics of the medical care.

Lines 371-372 (Discussions)

Another limitation is the descriptive nature of the analysis, without additional stratifications by treatments at baseline, stages, and histology types.

Reviewer 2

In this work, the authors described and analyzed real-world data from the subset data from the REFLECT study, pooled data from Central Eastern Europe (CEE), i.e., Bulgaria, Poland, Romania, Slovenia, for the treatment pattern and outcome of EGFR mutant patients. The authors concluded the importance of reflex testing and timely identification of patients and druggable mutations from initial diagnosis.

1.       Only 389 patients from 4 countries (CEE cohort)? That is really a small number.

2.       Table 1. From stage I to IV. There were 33 stage I, 13 stage II, and 19 stage IIIA, and 323 stage IV. Is there any Stage IIIB patient?

3.       The readers might be intriguing for the progression free and overall survival of different regimen or sequential regimen. Is there any analysis?

Thank you very much for the feedback provided. This is the first analysis of the population of patients with confirmed EGFR-mutated NSCLC receiving 1L 1G/2G EGFR TKIs in four Central Eastern European (CEE) countries. In this region, the registries are not yet uniformly implemented, and data collected in existing registries are not very standardized, making difficult any analysis of pooled data. Therefore, we considered participating in this study a great opportunity to be able to gather in a centralized manner data from this region.

Even though the number of patients included may be small (ie, n=389 at CEE level), it was a real challenge to build this population given due to a very restrictive study inclusion criterion, which did not allow prior systemic therapies for advanced / metastatic disease, other than EGFR TKIs. In this region, for example in medium-sized countries like Romania many patients were treated upfront with chemotherapy especially during 2015-2017, due to delayed authorization of innovative treatments from Health Insurance Houses. However, following your question, we revisited the manuscript and included a note on this topic. We have included in italic font in the column describing the changes the new and/or revisited text (lines 363-367).

Thank you very much for your observation captured in Question #2. There was an error in transcribing the stages in the brackets in this table. Now we have corrected the information on Table 1 (33 patients had stages I or II, 13 patients stage IIIA, 19 patients stage IIIB, and 323 patients stage IV). The corrections are highlighted in italic font in the manuscript changes summary column (Table 1).  

Finally, the REFLECT study was not powered to compare the individual 1G/2G EGFR TKIs. Therefore, outcomes could not be further characterized by molecule across the entire study, and we acknowledged this limitation in the manuscript.

Lines 363-365 (Discussions)

Yet, the number of patients was restricted considerably because no prior systemic therapy for advanced / metastatic disease was allowed, other than EGFR TKIs, and in some countries, chemotherapy was often used due to delays in EGFR TKIs authorizations by the Health Insurance Houses during 2015-2017. Due to these small numbers of patients, the generalizability of the results across countries or at country level is further hindered.

Table 1 – initial stage disease information

Reviewer 1

This is a pooled retrospective chart review of patients diagnosed with NSCLC to assess and describe treatment practices in TKI therapies. The paper is very clearly organized and well written; the tables and figures are appropriate and informative.

The primary objectives as stated were ..."the proportion of 1G/2G EGFR TKIs used as 1L therapy in patients with EGFRm NSCLC, real-world progression free survival (rwPFS) on 1L and treatment strategies in patients receiving 2L therapy." Figures 1 and 2 does not make sense with the stated objective. If the interest was to assess real world progression free survival then a Kaplan Meier plot would have been the most informative and in line with the objectives. I don't really understand the listing of the proportion blocks for describing this type of survival in any real informative way. Additionally, since multiple combinations of therapy were used it's difficult to discern how different therapies contribute to survival versus treatment changes. The authors should work with a statistician to design an analysis where this can be assessed or rewrite the objectives to just state that they are interested in describing treatment modalities by patient and tumor characteristics. The listed outcomes also make no sense with what is listed as the objective...as it is written the exposures are the combinations of the treatment modalities 1L, 2L (and 3L?) and the outcomes are death, progression, and treatment modality change. Although the figures are well done, it's difficult to discern any take home message from the data as presented. Please either exclude or separate patients that are on multiple treatments at baseline or censor them during follow-up when treatment modality changes. I have significant concerns about patient and treatment heterogeneity. Since the majority of patients are diagnosed at stage IV (83%) it makes sense to separate and focus on these or at the minimum do a sensitivity analysis where the cohort is separated into binary groupings: metastatic and non-metastatic. As 94% of tumors were adenocarcinomas, it also makes sense to do a focused analysis on this group since pathology and treatment can vary widely by histology. Was there additional testing done for mutation during follow-up? I would expect treatment guidelines to vary by clinic/country...please state whether this was considered or not.

Thank you very much for your detailed feedback and insightful recommendations.

Indeed, the REFLECT study aimed to provide as many insights as possible from the real-world practice, and it included multiple objectives. However, the main focus of the study was on patient attrition rates following first-line first- or second-generation (1L 1G/2G) EGFR TKIs. The study aimed to understand mainly the real-world sequencing, EGFR T790M testing rate and how the 1G/2G EGFR TKI therapy were being used in small and medium-sized countries from Europe, and to bring thus information from various geographies. In a region where registries are not uniformly implemented and/or data are collected in a less standardized manner, this is the first analysis of the population of patients with confirmed EGFR-mutated non-small cell lung cancer (NSCLC) receiving 1L 1G/2G EGFR TKIs in several Central Eastern European (CEE) countries, in a period when this was the standard treatment. Indeed, in this secondary manuscript we focused on treatments distribution and attrition rates across lines to build evidence on such outcomes for further referencing in this region, especially that the standard of care has changed, and data is more scarce compared to other regions or Western European countries. In addition, as you already remarked, differences across countries exist and due to this fact, we did not find relevant the Kaplan-Meier curve for the pooled data of CEE patients, especially when the primary manuscript of the study provided details by country (reference 24, Addeo A et al, Ther Adv Med Oncol. 2021). The objectives and outcomes have been defined at main study level, and the analyses have been conducted in a centralized way, which is currently limiting our possibility to address all suggestions in a timely manner fashion. However, building on your great feedback, we believe it may be more relevant to include additional details on rwPFS data by CEE country in the Results section and include a note on this in Discussions, to be thus more informative (Note: for consistency in reporting, we also described the overall survival data by country). We have included in italic font in the column describing the changes the new and/or revisited text (lines 211-219 and lines 336-339).

We are extremely thankful for the recommendations made and we will take them into account for further analyses in this cohort. For the time being, these are limitations; therefore, we have revisited this section and expanded on these topics. We have included in italic font in the column describing the changes the new and/or revisited text (lines 371-372).

The study collected mainly information on the T790M testing results; we gathered information on other mutations as well during initial EGFR testing and follow-up, but data were limited (ie, PD-L1 expression positive for <3% of patients, 1 patient with ALK rearrangement and 1 patient with TP53 variant mutations reported, whereas at the time of T790M testing, <1% patients were identified with PD-L1 positive expression and 1 patient with MET amplification).

Lines 211-219 (Results)

Median rwPFS on 1L EGFR TKI across the pooled CEE cohort was 14.0 (95% CI: 12.6; 15.6) months, with the following results at country level: Bulgaria 16.3 (95% CI: 11.5; 22.6) months, Poland 12.9 (95% CI: 10.3; 15.4) months, Romania 12.0 (95% CI: 10.3; 15.6) months and Slovenia 15.6 (95% CI: 12.6; 19.2) months, respectively.

At the end of data collection period, 225 (58%) deaths have been recorded. Median OS from the start of 1L EGFR TKI therapy was 26.6 (95% CI: 24.1; 29.0) months across the entire CEE cohort, with the following results by country: Bulgaria 23.4 (95% CI: 18.2;20.2) months, Poland 26.2 (95% CI: 18.0; 29.7) months, Romania 26.4 (95% CI 22.4; 34.2) months and Slovenia 28.9 (95% CI: 25.0; 34.3) months, respectively.

Lines 336 – 339 (Discussion)

However, rwPFS at country level in this CEE cohort varied, ranging from 12.0 months in Romania and 12.9 months in Poland to 15.6 months in Slovenia and 16.3 months in Bulgaria, differences probably explained by patients’ heterogeneity and local characteristics of the medical care.

Lines 371-372 (Discussions)

Another limitation is the descriptive nature of the analysis, without additional stratifications by treatments at baseline, stages, and histology types.

Reviewer 2

In this work, the authors described and analyzed real-world data from the subset data from the REFLECT study, pooled data from Central Eastern Europe (CEE), i.e., Bulgaria, Poland, Romania, Slovenia, for the treatment pattern and outcome of EGFR mutant patients. The authors concluded the importance of reflex testing and timely identification of patients and druggable mutations from initial diagnosis.

1.       Only 389 patients from 4 countries (CEE cohort)? That is really a small number.

2.       Table 1. From stage I to IV. There were 33 stage I, 13 stage II, and 19 stage IIIA, and 323 stage IV. Is there any Stage IIIB patient?

3.       The readers might be intriguing for the progression free and overall survival of different regimen or sequential regimen. Is there any analysis?

Thank you very much for the feedback provided. This is the first analysis of the population of patients with confirmed EGFR-mutated NSCLC receiving 1L 1G/2G EGFR TKIs in four Central Eastern European (CEE) countries. In this region, the registries are not yet uniformly implemented, and data collected in existing registries are not very standardized, making difficult any analysis of pooled data. Therefore, we considered participating in this study a great opportunity to be able to gather in a centralized manner data from this region.

Even though the number of patients included may be small (ie, n=389 at CEE level), it was a real challenge to build this population given due to a very restrictive study inclusion criterion, which did not allow prior systemic therapies for advanced / metastatic disease, other than EGFR TKIs. In this region, for example in medium-sized countries like Romania many patients were treated upfront with chemotherapy especially during 2015-2017, due to delayed authorization of innovative treatments from Health Insurance Houses. However, following your question, we revisited the manuscript and included a note on this topic. We have included in italic font in the column describing the changes the new and/or revisited text (lines 363-367).

Thank you very much for your observation captured in Question #2. There was an error in transcribing the stages in the brackets in this table. Now we have corrected the information on Table 1 (33 patients had stages I or II, 13 patients stage IIIA, 19 patients stage IIIB, and 323 patients stage IV). The corrections are highlighted in italic font in the manuscript changes summary column (Table 1).  

Finally, the REFLECT study was not powered to compare the individual 1G/2G EGFR TKIs. Therefore, outcomes could not be further characterized by molecule across the entire study, and we acknowledged this limitation in the manuscript.

Lines 363-365 (Discussions)

Yet, the number of patients was restricted considerably because no prior systemic therapy for advanced / metastatic disease was allowed, other than EGFR TKIs, and in some countries, chemotherapy was often used due to delays in EGFR TKIs authorizations by the Health Insurance Houses during 2015-2017. Due to these small numbers of patients, the generalizability of the results across countries or at country level is further hindered.

Table 1 – initial stage disease information

Reviewer 1

This is a pooled retrospective chart review of patients diagnosed with NSCLC to assess and describe treatment practices in TKI therapies. The paper is very clearly organized and well written; the tables and figures are appropriate and informative.

The primary objectives as stated were ..."the proportion of 1G/2G EGFR TKIs used as 1L therapy in patients with EGFRm NSCLC, real-world progression free survival (rwPFS) on 1L and treatment strategies in patients receiving 2L therapy." Figures 1 and 2 does not make sense with the stated objective. If the interest was to assess real world progression free survival then a Kaplan Meier plot would have been the most informative and in line with the objectives. I don't really understand the listing of the proportion blocks for describing this type of survival in any real informative way. Additionally, since multiple combinations of therapy were used it's difficult to discern how different therapies contribute to survival versus treatment changes. The authors should work with a statistician to design an analysis where this can be assessed or rewrite the objectives to just state that they are interested in describing treatment modalities by patient and tumor characteristics. The listed outcomes also make no sense with what is listed as the objective...as it is written the exposures are the combinations of the treatment modalities 1L, 2L (and 3L?) and the outcomes are death, progression, and treatment modality change. Although the figures are well done, it's difficult to discern any take home message from the data as presented. Please either exclude or separate patients that are on multiple treatments at baseline or censor them during follow-up when treatment modality changes. I have significant concerns about patient and treatment heterogeneity. Since the majority of patients are diagnosed at stage IV (83%) it makes sense to separate and focus on these or at the minimum do a sensitivity analysis where the cohort is separated into binary groupings: metastatic and non-metastatic. As 94% of tumors were adenocarcinomas, it also makes sense to do a focused analysis on this group since pathology and treatment can vary widely by histology. Was there additional testing done for mutation during follow-up? I would expect treatment guidelines to vary by clinic/country...please state whether this was considered or not.

Thank you very much for your detailed feedback and insightful recommendations.

Indeed, the REFLECT study aimed to provide as many insights as possible from the real-world practice, and it included multiple objectives. However, the main focus of the study was on patient attrition rates following first-line first- or second-generation (1L 1G/2G) EGFR TKIs. The study aimed to understand mainly the real-world sequencing, EGFR T790M testing rate and how the 1G/2G EGFR TKI therapy were being used in small and medium-sized countries from Europe, and to bring thus information from various geographies. In a region where registries are not uniformly implemented and/or data are collected in a less standardized manner, this is the first analysis of the population of patients with confirmed EGFR-mutated non-small cell lung cancer (NSCLC) receiving 1L 1G/2G EGFR TKIs in several Central Eastern European (CEE) countries, in a period when this was the standard treatment. Indeed, in this secondary manuscript we focused on treatments distribution and attrition rates across lines to build evidence on such outcomes for further referencing in this region, especially that the standard of care has changed, and data is more scarce compared to other regions or Western European countries. In addition, as you already remarked, differences across countries exist and due to this fact, we did not find relevant the Kaplan-Meier curve for the pooled data of CEE patients, especially when the primary manuscript of the study provided details by country (reference 24, Addeo A et al, Ther Adv Med Oncol. 2021). The objectives and outcomes have been defined at main study level, and the analyses have been conducted in a centralized way, which is currently limiting our possibility to address all suggestions in a timely manner fashion. However, building on your great feedback, we believe it may be more relevant to include additional details on rwPFS data by CEE country in the Results section and include a note on this in Discussions, to be thus more informative (Note: for consistency in reporting, we also described the overall survival data by country). We have included in italic font in the column describing the changes the new and/or revisited text (lines 211-219 and lines 336-339).

We are extremely thankful for the recommendations made and we will take them into account for further analyses in this cohort. For the time being, these are limitations; therefore, we have revisited this section and expanded on these topics. We have included in italic font in the column describing the changes the new and/or revisited text (lines 371-372).

The study collected mainly information on the T790M testing results; we gathered information on other mutations as well during initial EGFR testing and follow-up, but data were limited (ie, PD-L1 expression positive for <3% of patients, 1 patient with ALK rearrangement and 1 patient with TP53 variant mutations reported, whereas at the time of T790M testing, <1% patients were identified with PD-L1 positive expression and 1 patient with MET amplification).

Lines 211-219 (Results)

Median rwPFS on 1L EGFR TKI across the pooled CEE cohort was 14.0 (95% CI: 12.6; 15.6) months, with the following results at country level: Bulgaria 16.3 (95% CI: 11.5; 22.6) months, Poland 12.9 (95% CI: 10.3; 15.4) months, Romania 12.0 (95% CI: 10.3; 15.6) months and Slovenia 15.6 (95% CI: 12.6; 19.2) months, respectively.

At the end of data collection period, 225 (58%) deaths have been recorded. Median OS from the start of 1L EGFR TKI therapy was 26.6 (95% CI: 24.1; 29.0) months across the entire CEE cohort, with the following results by country: Bulgaria 23.4 (95% CI: 18.2;20.2) months, Poland 26.2 (95% CI: 18.0; 29.7) months, Romania 26.4 (95% CI 22.4; 34.2) months and Slovenia 28.9 (95% CI: 25.0; 34.3) months, respectively.

Lines 336 – 339 (Discussion)

However, rwPFS at country level in this CEE cohort varied, ranging from 12.0 months in Romania and 12.9 months in Poland to 15.6 months in Slovenia and 16.3 months in Bulgaria, differences probably explained by patients’ heterogeneity and local characteristics of the medical care.

Lines 371-372 (Discussions)

Another limitation is the descriptive nature of the analysis, without additional stratifications by treatments at baseline, stages, and histology types.

Reviewer 2

In this work, the authors described and analyzed real-world data from the subset data from the REFLECT study, pooled data from Central Eastern Europe (CEE), i.e., Bulgaria, Poland, Romania, Slovenia, for the treatment pattern and outcome of EGFR mutant patients. The authors concluded the importance of reflex testing and timely identification of patients and druggable mutations from initial diagnosis.

1.       Only 389 patients from 4 countries (CEE cohort)? That is really a small number.

2.       Table 1. From stage I to IV. There were 33 stage I, 13 stage II, and 19 stage IIIA, and 323 stage IV. Is there any Stage IIIB patient?

3.       The readers might be intriguing for the progression free and overall survival of different regimen or sequential regimen. Is there any analysis?

Thank you very much for the feedback provided. This is the first analysis of the population of patients with confirmed EGFR-mutated NSCLC receiving 1L 1G/2G EGFR TKIs in four Central Eastern European (CEE) countries. In this region, the registries are not yet uniformly implemented, and data collected in existing registries are not very standardized, making difficult any analysis of pooled data. Therefore, we considered participating in this study a great opportunity to be able to gather in a centralized manner data from this region.

Even though the number of patients included may be small (ie, n=389 at CEE level), it was a real challenge to build this population given due to a very restrictive study inclusion criterion, which did not allow prior systemic therapies for advanced / metastatic disease, other than EGFR TKIs. In this region, for example in medium-sized countries like Romania many patients were treated upfront with chemotherapy especially during 2015-2017, due to delayed authorization of innovative treatments from Health Insurance Houses. However, following your question, we revisited the manuscript and included a note on this topic. We have included in italic font in the column describing the changes the new and/or revisited text (lines 363-367).

Thank you very much for your observation captured in Question #2. There was an error in transcribing the stages in the brackets in this table. Now we have corrected the information on Table 1 (33 patients had stages I or II, 13 patients stage IIIA, 19 patients stage IIIB, and 323 patients stage IV). The corrections are highlighted in italic font in the manuscript changes summary column (Table 1).  

Finally, the REFLECT study was not powered to compare the individual 1G/2G EGFR TKIs. Therefore, outcomes could not be further characterized by molecule across the entire study, and we acknowledged this limitation in the manuscript.

Lines 363-365 (Discussions)

Yet, the number of patients was restricted considerably because no prior systemic therapy for advanced / metastatic disease was allowed, other than EGFR TKIs, and in some countries, chemotherapy was often used due to delays in EGFR TKIs authorizations by the Health Insurance Houses during 2015-2017. Due to these small numbers of patients, the generalizability of the results across countries or at country level is further hindered.

Table 1 – initial stage disease information