Gadgeel Discusses Therapies for Patients With NSCLC and CNS Metastases

Shirish M. Gadgeel, MD

Division Head for Hematology/Oncology

Associate Director, Patient Experience and Clinical Care

Henry Ford Health System

Detroit, MI

Targeted Oncology^TM: Can you review the recommendations for molecular testing for patients who may be candidates for these agents?

GADGEEL: The NCCN [National Comprehensive Cancer Network] recommends [obtaining] the biomarkers before administering first-line ICI [immune checkpoint inhibitor] therapy in nonsquamous non–small cell lung cancer [NSCLC].¹

They don’t say first-line therapy because there are clearly [cases in which] you are compelled to start therapy before the molecular test results are available. But if they are not available and you feel a clinical need to start treatment, you can start systemic chemotherapy but avoid starting immunotherapy even if the PD-L1 is high.

Although the PD-L1 expression can be elevated in patients with oncogenic-driven NSCLC, targeted therapy for the oncogenic driver should always take precedence over treatment with an ICI. In patients with PD-L1 expression greater than 1% with a targetable driver alteration, [patients] should receive first-line targeted therapy for that oncogene and not first-line ICI, because targeted therapy yields high response rates in the first-line setting.

The first point is ensuring that you know what tests your NGS [next-generation sequencing] lab is doing; ensure that RNA-based testing is being performed to capture all the fusions. The second point is, try to avoid starting patients on immunotherapy until you have the molecular results back.

Which trials in nonsquamous NSCLC relate to this patient case? Which efficacy end points are most meaningful in these studies?

I’m going to be reviewing the clinical trial data for 3 studies: ALEX [NCT02075840], ALTA-1L [NCT02737501], and CROWN [NCT03052608], which led to the approval of [alectinib (Alecensa), brigatinib (Alunbrig), and lorlatinib (Lorbrena)], respectively.^2-4

It is always difficult to do cross-trial comparison, but…we must make the best decision. The efficacy end points that we look for are the PFS [progression-free survival], the HR [hazard ratio], and the intracranial responses, as well as the intracranial PFS.

We look for the overall efficacy of the drug but in these patients—particularly because they are young and can live for a long time and therefore can develop brain metastases—the brain activity is critical. Each of these drugs has [brain activity] but there are some differences.

What are the key takeaways from the ALEX study?

ALK was tested centrally, so you had to submit tissue and they had to centrally confirm this was ALK-positive [cancer]—only then could you randomize patients.² Whenever you must test centrally, there’s a little bit of a selection bias. Nonetheless, these patients were randomized to alectinib vs crizotinib [Xalkori]. Like all the other trials, it allowed asymptomatic brain metastases, and PFS was the end point. In this study, it was investigator assessed.

In the final analysis [of PFS, there was] an HR of 0.43 [95% CI, 0.32-0.58; P < .001]. The median PFS was 35 months with alectinib, almost 3 years. This was investigator assessed and it is 1 time point in the entire curve. I prefer to look at the HR, because it gives you an idea over the entire course of the 2 Kaplan-Meier curves. It’s an easier number to compare across trials.

[Let’s discuss] the PFS according to central nervous system [CNS] metastases of both alectinib and crizotinib. Patients with CNS metastases, in general, tend to do a bit worse with these drugs than patients without CNS metastases at baseline. At 4 years in patients without CNS metastases [receiving alectinib], 47.8% of the patients were progression free, whereas that percentage was 38% with alectinib [in those who had CNS metastases].

Then you look at crizotinib; by the time you reach even 1 year, the PFS rate in patients with CNS metastases is only about 30%. There’s no question that these next-generation drugs are better at CNS control than crizotinib. But we do need to remember that patients who have baseline CNS metastases do end up doing a bit worse than patients without CNS metastases.

The [data for 5-year overall survival (OS) are] remarkable, although not statistically significant. The HR was 0.67, with a P value of .03, but it was not statistically significant because not enough events had occurred [95% CI, 0.46-0.98]. The 5-year survival rate was 62% with alectinib and 45% with crizotinib—17 percentage point difference in survival at 5 years.

I want to put this into perspective. Some folks have raised the concern that we still don’t see any survival advantage; what’s the harm of starting the patient on crizotinib first and then going to alectinib? To provide perspective, if you looked at the FLAURA EGFR trial [NCT02296125], the median survival in patients who got osimertinib was 38 months.⁵

Even in the crizotinib arm, patients’ 5-year survival was 45%, whereas the median survival with osimertinib was 38 months. You can see that whenever there’s a longer survival, it takes much longer follow-up to show survival advantage. Even though a survival advantage has not been seen, the trend is clearly in favor of using the next-generation drug over starting with crizotinib. The OS according to the different subgroups [showed that] in general, most of the groups did better, although the interaction test was not statistically significant.² What I would like to highlight is the HR with patients with CNS metastases.

This is not PFS, this is OS. One needs to be mindful of that. When you look at CNS metastases, the hazard ratio is 0.58; there’s almost a 50% better survival in patients with CNS metastases with alectinib compared with crizotinib [95% CI, 0.34-1.00; P = .47].

What efficacy data are important to understand from the ALTA-1L trial?

The other drug that is approved now is brigatinib, [which was studied in the ALTA-1L trial]. This was similar to the ALEX study but one important difference was [that] this allowed local test results.^2,3

I think that this study is a little more real world than ALEX because it allowed patients to be enrolled based on local result.³ I do believe when these studies were conducted, we did not have the precision that we have of identifying ALK translocation as we have now, with NGS. With immunohistochemistry and fluorescence in situ hybridization, there is a possibility of misdiagnosis. One needs to take that into consideration. Another important point is [the trial investigators] did allow 1 cycle of chemotherapy if molecular results were not available, and they became available afterward. Those are 2 main differences between ALTA-1L and ALEX.^2,3

The updated PFS data [in ALTA-1L had an HR of 0.43 (95% CI, 0.31-0.61; P < .0001)] in the investigator-assessed PFS, [which was] very identical to what was seen in ALEX.³ The median PFS was 29 months, as compared with 9 months with crizotinib. The numbers are a little bit lower than what was seen in ALEX for both arms, but the HR is identical. So, very similar efficacy between brigatinib and alectinib, at least when you do a cross-trial comparison—something that they tell us not to do but we always end up doing.

In the subgroup analysis [of PFS for ALTA-1L, in those with] brain metastases the PFS hazard ratio was 0.25 [95% CI, 0.14-0.46]. One of the biggest areas where the next-generation drugs have been very beneficial are in patients with CNS metastases. Not only do patients with baseline brain metastases do better with these next-generation drugs, but also it appears that [the drugs] delay the onset of development of brain metastases in patients who don’t have brain metastases at baseline.

They appear to have a CNS protective effect. The CNS control is an important component of why these drugs provide better benefit than crizotinib, like we saw for ALEX. Control is better in patients with baseline brain metastases, with an HR of 0.25 compared with patients without brain metastases at 0.65 [95% CI, 0.44-0.97; P = .03].

The response rate was 81% with brigatinib and 73% with crizotinib. When you look at the response rate, something that you measure within 6 to 12 weeks, the response rates are not that different between these next-generation ALK inhibitors and crizotinib.

The difference emerges in that the disease control, or the response duration, is far longer with these drugs than with crizotinib. [For example, the] median duration of confirmed response was not reached with brigatinib and was 13.8 months [with crizotinib]. What is critical is that the overall response rate is not that much better, but the duration of control and duration of response are far longer with these drugs than with crizotinib.

That’s what leads to the improved PFS. Sometimes it is challenging to assess brain responses, because not all the brain metastases may be measurable, and some of these patients did get stereotactic radiosurgery.

It becomes difficult to measure those lesions. But the response rate was 78% in the brigatinib patients and 26% in the crizotinib patients [odds ratio (OR), 11.67; 95% CI, 2.15-63.27; P = .0014]. If you include patients with nonmeasurable brain metastases, it was [respectively] 66% and 16% [OR, 11.75; 95% CI, 4.19-32.91; P < .0001]. We don’t have as much of a follow-up within this study as compared with the ALEX trial for OS [HR, 0.92; 95% CI, 0.57-1.47; P = .772]. It is important to note that 44% of the patients in the crizotinib arm crossed over to brigatinib, and in these patients the PFS with brigatinib was 15.6 months.

Patients do well, but the PFS with brigatinib is not the same as was seen with brigatinib in the frontline setting. So, important to note that in the second-line setting, although 15 months is quite impressive, it’s not the same as what was seen with brigatinib in the frontline setting.

How did patients with NSCLC do in the CROWN study?

The CROWN trial was presented and published last year. This was a study that randomized patients to lorlatinib vs crizotinib.⁴ It allowed asymptomatic brain metastases. This was centrally tested ALK and the primary end point was PFS by independent review. ALEX was investigator reviewed, ALTA-1L was independent reviewed, and CROWN was also independent reviewed. [PFS in CROWN is] one of the most impressive results. We don’t have as much follow-up, so one needs to be mindful; I keep talking about being careful about cross-trial comparisons. But the HR here was 0.28 [95% CI, 0.19-0.41; P < 0.001]. Just to remind you, the independent review HR in ALTA-1L and ALEX was 0.48 [95% CI, 0.35-0.66] and 0.43 [95% CI, 0.32-0.58], respectively; so, 0.28 appears very impressive, and appears much more pronounced than what was seen in ALEX or in ALTA-1L, and that benefit was seen across all subgroups.

The confirmed response rate was 76% [OR, 2.25; 95% CI, 1.35-3.89]. In ALEX and ALTA-1L, the numbers were about 80% and that was not a confirmed response rate.^2,3 This is confirmed response rate, and this was 76% with lorlatinib, and crizotinib was 58%.⁴

Now, this appears a little more different, but please note that this difference is not as pronounced as it is with the PFS, and it’s primarily because…when you look at median DOR [duration of response], it was not evaluable in the lorlatinib arm, and it was 11 months with crizotinib. Again, we looked at what the response rate was in patients with brain metastases. If you look at both measurable and nonmeasurable metastases, the response rate was 66%. It was 20% with crizotinib. Those numbers—and this is confirmed response rate—in patients with measurable disease was 82% and crizotinib, 23%. The Kaplan-Meier curves showing the intracranial PFS with lorlatinib as compared with crizotinib [in the CROWN study] almost appears as a flat line. The follow-up is short, and one needs to view that with a level of caution. But the CNS activity, coupled with the HR of 0.28 with lorlatinib, I’ve started considering lorlatinib in patients, particularly with CNS metastases.

As I mentioned, the patients with CNS metastases don’t do as well as patients without brain metastases at baseline; and because of the activity of lorlatinib in those patients with brain metastases, I’ve started considering [it]. I’m not saying I use it in every patient, but I’ve considered lorlatinib because I feel that the CNS activity of this drug is quite impressive.

What can you tell us about the safety data for lorlatinib, especially cognitive effects?

There are clearly issues with cognitive effects.⁴ They can be different things, they can get forgetful, their personalities may change. This is quite relevant because these are generally individuals in the prime of their careers, they’re in their early 50s, and so that may have more of an impact. Because of that, when I do start lorlatinib either as first-line or second-line treatment, I often will start them on the 25-mg pills and ask them to take 4 pills, because I can adjust the dose much more easily. I can come down to a 75-mg dose or 50 mg, instead of starting with the 100-mg pills, because then you must go back to ordering the 25-mg pills. I start off with 25-mg pills so I can adjust the dose if any of those adverse events [AEs] occur, and my threshold to lower the dose to 75 mg or 50 mg is quite low. Generally, these AEs with lorlatinib reverse quickly once you reduce the dose.

How do these 3 trials compare?

The median follow-up is clearly longest with ALEX at 37.8 months.² It was 29.4 months with ALTA and only 18.0 months with CROWN, and that needs to be [considered].^3,4 But if you look at the median PFS, which was the primary end point, it was 34.8 months with ALEX; that was the investigator-assessed end point, so that was the primary end point.²

In ALTA-1L, it was independent review [and was] 24.0 months, but if you look at investigator-assessed PFS, it was about 29.4 months.³ So, much more comparable to ALEX.^3,4 And the HRs were very comparable as well as with ALTA-1L with an investigator-assessed PFS, but when it came to CROWN, the PFS is not reached, at least at the 18.3-month follow-up.⁴

If you look at the intracranial overall response rate, they don’t look much different, but what is impressive is the median brain PFS, and it appears the most promising is in CROWN.^2-4

What can you tell us about the similarities and differences for the safety data for these agents?

These are the safety issues. The notable AEs are anemia, [elevated] liver function tests, peripheral edema, myalgias, increased weight, and sun sensitivity.^2-4,6 I would like to emphasize [that] the sun sensitivity with alectinib happens.^2,6

I’m in Michigan, so sun is not something we have the privilege of seeing on a regular basis, but folks who are more in the southern states need to be mindful about that. With ALTA-1L [there] is the issue of early pulmonary toxicity.³ As we have discussed with CROWN, with lorlatinib, there is the issue of the cognitive effects.⁴ I would also like to highlight that the edema appears much more pronounced with lorlatinib. And there’s, of course, the issue of hypercholesterolemia and hypertriglyceridemia. There are warnings and precautions listed in the package insert [for lorlatinib].⁷ Generally, when it has come to alectinib and brigatinib, I have not seen these come up, except for the myalgias and the blood count AST [aspartate aminotransferase]/ALT [alanine aminotransferase] changes. And with ALTA-1L, similarly. So yes, there are warnings, and one needs to be mindful of that, but usually [these safety issues] are not seen.

_REFERENCES

_{1. NCCN. Clinical Practice Guidelines in Oncology. Non-small cell lung cancer, version 1.2022. Accessed January 11, 2022. https://bit.ly/33qNCn6}

_{2. Mok T, Camidge DR, Gadgeel SM, et al. Updated overall survival and final progression-free survival data for patients with treatment-naive advanced ALK-positive non-small-cell lung cancer in the ALEX study. Ann Oncol. 2020;31(8):1056-1064. doi:10.1016/j.annonc.2020.04.478}

_{3. Camidge DR, Kim HR, Ahn MJ, et al. Brigatinib versus crizotinib in advanced ALK inhibitor-naive ALK-positive non-small cell lung cancer: second interim analysis of the phase III ALTA-1L trial. J Clin Oncol. 2020;38(31):3592-3603. doi:10.1200/JCO.20.00505}

_{4. Shaw AT, Bauer TM, de Marinis F, et al; CROWN Trial Investigators. First-line lorlatinib or crizotinib in advanced ALK-positive lung cancer. N Engl J Med. 2020;383(21):2018-2029. doi:10.1056/NEJMoa2027187}

_{5. Ramalingam SS, Vansteenkiste J, Planchard D, et al; FLAURA Investigators. Overall survival with osimertinib in untreated, EGFR-mutated advanced NSCLC. N Engl J Med. 2020;382(1):41-50. doi:10.1056/NEJMoa1913662}

_{6. Peters S, Camidge DR, Shaw AT, et al; ALEX Trial Investigators. Alectinib versus crizotinib in untreated ALK-positive non-small-cell lung cancer. N Engl J Med. 2017;377(9):829-838. doi:10.1056/NEJMoa1704795}

_{7. Lorbrena. Prescribing information. Pfizer Laboratories; 2021. Accessed January 11, 2022. https://bit.ly/3fooX5d}