Palliative Systemic Therapy Given near the End of Life for Metastatic Non-Small Cell Lung Cancer

Round 1

Reviewer 1 Report

Abstract (L16) - SACT - undefined abbreviation.

Introduction (L68) - "has"  should be "have" - grammar

Methods (L89+) - Please state here that the study was approved by an appropriate Research Ethics Committee. I do appreciate that it is already provided in the footnotes.

Study design issue - Why was data collection truncated to the end of October 2016?  There must have been many more patients since then.  This issue is relevant to the study limitations declared in the Discussion section of the manuscript.

Results - Table 1 - Provide age to integer accuracy only.  Higher accuracy is clinically irrelevant.

Section 3.2 - (L129-30) -Redundant text presentation with Table 2

(L138-9) - Leave out mean and SD.  These data are not normally distributed,

and median and IQR are the appropriate descriptors.

Don't repeat data already presented in Table 3 in the text.

L159-62 - Don't repeat data from Fig 2 and its legend in the text.   This issue also applies to L168-70 with respect to Fig 3.

Discussion -  Don't repeat results here.  Only integrate your findings into the published scientific literature on this subject.

Author Response

Reviewer’ comments:

We thank the reviewers for their constructive comments, which helped improve our manuscript and were all addressed.

Reviewer #1:

1. Abstract (L16) - SACT - undefined abbreviation.
   • We added the definition (systemic anticancer therapy).
2. Introduction (L68) - "has" should be "have" – grammar
   • We made the correction.
3. Methods (L89+) - Please state here that the study was approved by an appropriate Research Ethics Committee. I do appreciate that it is already provided in the footnotes.
   • This has been added in the text as well.
4. Study design issue - Why was data collection truncated to the end of October 2016? There must have been many more patients since then.  This issue is relevant to the study limitations declared in the Discussion section of the manuscript.
   • The purpose was the perform a study shortly after the approval of new therapies for NSCLC (second-line nivolumab and pembrolizumab). There were delays for collecting data, statistical analysis, and then lack of time during the pandemy to submit the manuscript.

 

5. Results - Table 1 - Provide age to integer accuracy only. Higher accuracy is clinically irrelevant.

• We made the correction.

6. Section 3.2 - (L129-30) -Redundant text presentation with Table 2
   • We made the correction.
7. (L138-9) - Leave out mean and SD. These data are not normally distributed, and median and IQR are the appropriate descriptors.
   • We made the correction.
8. Don't repeat data already presented in Table 3 in the text.
   • We removed numbers and p values, but we kept the explanations to help the reader understand the table and to emphasize the important findings.
9. L159-62 - Don't repeat data from Fig 2 and its legend in the text. This issue also applies to L168-70 with respect to Fig 3.
   • We made the corrections for both situations.
10. Discussion - Don't repeat results here.  Only integrate your findings into the published scientific literature on this subject.
    • We removed most of the results that were repeated in the discussion.

 

Author Response File: Author Response.docx

Reviewer 2 Report

This is a single centre, retrospective cohort study examining the use of systemic anticancer therapies (SACT) within 30 days of death in patients with metastatic non-small cell lung cancer (NSCLC).  In the cohort of 90 eligible patients who died between November 1, 2014 and October 31, 2016, the mean delay between last treatment and death was 94 days (SD 107 days) and the median was 57 days (IQR 35-103 days). Twenty patients (22%) received SACT within 30 days of death. One of these patients died from treatment-related toxicity. Demographic and tumour-related characteristics were similar amongst those who did and did not receive SACT within 30 days of death although there were some trends to note (discussed below). Those who did not receive SACT near the end of life were more likely to be seen by palliative care, have MaiD or palliative sedation, and die in hospice or at home as opposed to in hospital. Of note, 77% of patients who were not treated near the end of life died in hospital in comparison to 85% who received SACT within 30 days of death. Adjusted median OS was 9.0 vs 4.0 months for those who were not treated near the end of life compared with those received SACT within 30 days of death.  

This was a well-organized and easy to follow manuscript. As highlighted, this study fills a gap as it has documented SACT use near the end of life in the more modern era of therapies including TKIs and immunotherapy in addition to chemotherapy. Additionally, this study provided rich data on tumour characteristics, therapies/number of cycles received by line of treatment, detailed cause of death and more. It is unlikely that a large, multicentre study would have been able to achieve such granular data. In a Canadian/public-payer context, results can be used for benchmarking of quality EOL outcomes amongst patients treated with SACT for metastatic NSCLC.

Further comments/queries:

1. Excellent introduction
2. Line 68: “has” should be “were”
3. Patients were identified through an Oncology Database (SICTO). Please provide more information about this database and reference if appropriate. Was the demographic, tumour-related, treatment, and outcomes data abstracted from SICTO as well or was this through chart-review? Please state.
4. Excellent to include a patient-flow diagram for accrual into a retrospective study
5. Levels of care: Are there only 2 levels in Quebec?
6. Figure 2/Figure 3: Figure 3 could be deleted. Information provided in text is adequate.
7. Could you describe or have a table showing distribution of treatments received within 30 days of death for this group that received SACT near end of life?
8. Discussion: Line 176: “This study being retrospective, it was not easy to understand why these patients were treated near end of life…” Although there were not statistical differences in the two groups, it would be important to discuss some trends that could be explanatory. I wonder if there could have been two types of patients getting treated near end of life. Proportionally, those treated near end of life included more patients with EGFR mutations and ALK rearrangements (and more of these patients received TKIs). In comparison to IO/chemotherapy, TKIs are easy to continue despite clinical deterioration. This is not an excuse for continuation just a potential explanation. The second group who I wonder might be more likely to receive SACT near end of life would be those with more aggressive disease. Clinical deterioration can often take the patient and care team by surprise. A clue for this in your data is that proportionally, there were more patients who died during or just after the first line of treatment in the SACT near end of life group. Less patients in this group made it to 3^rd line or beyond. Although SACT within the last 30 days of life is an agreed upon quality metric for EOL care in cancer and should still be minimized and measured, I suspect that we will never see no patients falling into this category due to some cancers just being more aggressive and leading to unexpected/rapid deterioration.
9. Paragraph starting on line 193: Please discuss overall rate of death in hospital in comparison to literature for metastatic NSCLC +/- other tumour types. Even the 77% for those who did not have SACT at the end of life seems high so context would be helpful for the reader.
10. Sentence starting line 204: Expand on other possible factors. Related to discussion above, at least some of these patients may have had more aggressive/less treatment-responsive disease (biology).

Author Response

Reviewer’ comments:

We thank the reviewers for their constructive comments, which helped improve our manuscript and were all addressed.

Reviewer #2:

1. Excellent introduction
   • Thank you.
2. Line 68: “has” should be “were”
   • We made the correction.
3. Patients were identified through an Oncology Database (SICTO). Please provide more information about this database and reference if appropriate. Was the demographic, tumour-related, treatment, and outcomes data abstracted from SICTO as well or was this through chart-review? Please state.
   • We added the information.
4. Excellent to include a patient-flow diagram for accrual into a retrospective study
   • Thank you.
5. Levels of care: Are there only 2 levels in Quebec?
   • Now there are 4 (A-B-C-D). In 2016, there used to be 3, but level 3 corresponded to comfort care, so no patient receiving SACT would have been level 3. We added the definition of level 3 in Table 3 legend.
6. Figure 2/Figure 3: Figure 3 could be deleted. Information provided in text is adequate.
   • We kept Figure 3, but we removed duplicate results from the text as suggested by reviewer #1.
7. Could you describe or have a table showing distribution of treatments received within 30 days of death for this group that received SACT near end of life?
   • The information has been added to Table 3.
8. Discussion: Line 176: “This study being retrospective, it was not easy to understand why these patients were treated near end of life…” Although there were not statistical differences in the two groups, it would be important to discuss some trends that could be explanatory. I wonder if there could have been two types of patients getting treated near end of life. Proportionally, those treated near end of life included more patients with EGFR mutations and ALK rearrangements (and more of these patients received TKIs). In comparison to IO/chemotherapy, TKIs are easy to continue despite clinical deterioration. This is not an excuse for continuation just a potential explanation. The second group who I wonder might be more likely to receive SACT near end of life would be those with more aggressive disease. Clinical deterioration can often take the patient and care team by surprise. A clue for this in your data is that proportionally, there were more patients who died during or just after the first line of treatment in the SACT near end of life group. Less patients in this group made it to 3rd line or beyond. Although SACT within the last 30 days of life is an agreed upon quality metric for EOL care in cancer and should still be minimized and measured, I suspect that we will never see no patients falling into this category due to some cancers just being more aggressive and leading to unexpected/rapid deterioration.

• Thank you for these thoughtful comments. The discussion has been expanded to take these comments into account.

9. Paragraph starting on line 193: Please discuss overall rate of death in hospital in comparison to literature for metastatic NSCLC +/- other tumour types. Even the 77% for those who did not have SACT at the end of life seems high so context would be helpful for the reader.
   • We expanded the discussion to answer your comment.
10. Sentence starting line 204: Expand on other possible factors. Related to discussion above, at least some of these patients may have had more aggressive/less treatment-responsive disease (biology).
    • We expanded earlier in the discussion as per your comment #8.

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

Table 1 - Mean age (all)   S.D. should be truncated to integer accuracy, i.e. "7".

All of my other issues are satisfactorily addressed.