Later-Line Treatment with Lorlatinib in ALK- and ROS1-Rearrangement-Positive NSCLC: A Retrospective, Multicenter Analysis

In clinical practice, patients with anaplastic lymphoma kinase (ALK)-rearrangement–positive non–small-cell lung cancer commonly receive sequential treatment with ALK tyrosine kinase inhibitors. The third-generation agent lorlatinib has been shown to inhibit a wide range of ALK resistance mutations and thus offers potential benefit in later lines, although real-world data are lacking. This multicenter study retrospectively investigated later-line, real-world use of lorlatinib in patients with advanced ALK- or ROS1-positive lung cancer. Fifty-one patients registered in a compassionate use program in Austria, who received second- or later-line lorlatinib between January 2016 and May 2020, were included in this retrospective real-world data analysis. Median follow-up was 25.3 months. Median time of lorlatinib treatment was 4.4 months for ALK-positive and 12.2 months for ROS-positive patients. ALK-positive patients showed a response rate of 43.2%, while 85.7% percent of the ROS1-positive patients were considered responders. Median overall survival from lorlatinib initiation was 10.2 and 20.0 months for the ALK- and ROS1-positive groups, respectively. In the ALK-positive group, lorlatinib proved efficacy after both brigatinib and alectinib. Lorlatinib treatment was well tolerated. Later-line lorlatinib treatment can induce sustained responses in patients with advanced ALK- and ROS1-positive lung cancer.

be difficult or even impossible in the advanced setting, because ALK-positive tumors tend to metastasize to the CNS. Liquid biopsy, which is well established in EGFR-mutant disease, is presently under validation for ALK-positive NSCLC [22]. The National Comprehensive Cancer Network NCCN [23] and International Association for the Study of Lung Cancer/College of American Pathologists/ Association for Molecular Pathology (IASCL/CAP/AMP) guidelines [24] do not recommend ALK mutation testing, while the European Society for Medical Oncology (ESMO) guidelines [25] recommend testing but add that this is not mandatory for treatment decisions.
Lorlatinib is a potent, CNS-active, highly selective ALK/ROS1 inhibitor that has been approved in the European Union, the US, Canada and Japan as a monotherapy for the treatment of adult patients with ALK-positive advanced NSCLC, whose disease has progressed after alectinib or ceritinib as the first ALK TKI, or crizotinib and at least one other ALK TKI [26,27]. In the pivotal phase II trial, lorlatinib has shown substantial overall and intracranial activity [28]. Real-world data on the use of this agent are lacking, however. Therefore, we retrospectively assessed the efficacy and tolerability of lorlatinib in patients with ALK-positive or ROS1-positive lung cancer.

ALK-Positive Patients
The median follow-up of the ALK-positive patients was 24.8 months. At the time of analysis, 11 (29.7%) out of the 37 ALK-positive patients were still on lorlatinib treatment.
Our patients showed a median duration on treatment (DoT) of 4.4 months (95% CI: 1.3; 7.6). (Table 2, Figure 1) Among the subgroups, the patients who received one line of prior ALK-TKI had a DoT of 4.4 months (n = 10; 95% CI: 0.5; 8.2) compared to 3.0 months (n = 14; 95% CI: 1.8; 4.2) in the group that received ≥ 3 prior lines of TKI with no significant difference between the subgroups. The subgroup with two prior lines of TKI has not yet reached a discontinuation rate of 50%.
The ORR for patients who received only one prior TKI was 40.0% (95% CI: 12.2; 73.8) compared to those with two prior lines of TKI who had an ORR of 53.8% (95% CI: 25.1; 80.8), on the contrary to the patients who received three or more lines of TKI. Those had an ORR of 35.7% (95% CI: 12.8; 64.9). No significant difference was observed ( Table 2).
Eight patients had confirmed brain metastasis at initiation of lorlatinib and no following cranial radiotherapy. Out of these patients suitable for providing data for intracranial response, five (62.5%) showed an intracranial PR (1 stable disease (SD), 2 progressive disease (PD)).
Patients who received only one second generation TKI prior to lorlatinib showed a comparable survival time since date of locally advanced or metastatic diagnosis, with 39.2 months (95% CI: 21.7; 56.7) compared to 43.4 months (95% CI: 30.9; 55.9) in patients who received more than one second generation TKI. There was, significantly, no difference in survival (p = 0.003) between these groups ( Table 2).

ROS1-Positive Patients
Out of the 14 ROS1-positive patients, 6 patients (42.9%) were still on treatment and a median follow-up time of 32.2 months was reached.
The median DoT of all patients was 12.2 months (95% CI: 0.0; 29.5) with a range of 1.0 to 41.6 months of treatment. Due to the lower number of patients and the homogeneity regarding the prior therapies, no subgroup analysis was conducted (Table 2, Figure 2).

Tolerability
Forty-nine patients received lorlatinib at the approved starting dose of 100 mg/day. Due to fears of toxicities, two patients were given 50 mg/day for one week before increasing to the standard dose. In both cases, the full dose was well tolerated. Dose reductions due to adverse events were necessary for 9 of the 51 patients (18%), with lorlatinib 25 mg/day as the lowest daily dose used.

Discussion
The optimal treatment sequence in the setting of ALK-positive NSCLC is of substantial clinical relevance. ALK TKIs are often particularly effective in these patients, thus changing the course of their disease, and their optimal sequence might affect patient outcomes to a considerable extent. As controlled clinical trials investigating various sequences of ALK TKIs across multiple treatment lines will presumably never become available, real-world data can add important information to the body of evidence. For instance, a retrospective analysis conducted across 23 Italian centers showed that sequential administration of crizotinib and a next-generation ALK TKI provided remarkable clinical benefits in a real-life population of 290 patients [29].
Lorlatinib appears to be an ideal choice in the later-line setting, given its uniquely broad activity against ALK resistance mutations. According to in-vitro data, most of the commonly observed secondary ALK mutations are inhibited by lorlatinib [13]. The pivotal clinical trial demonstrated rapid, deep and durable responses with lorlatinib after up to three prior ALK TKIs that had been administered with or without chemotherapy [21,28]. Furthermore, it has shown an improvement in global quality of life for the majority of patients treated with lorlatinib [30]. Hypercholesterolemia and hypertriglyceridemia represented the most common treatment-related AEs, and the antitumor activity of lorlatinib was seen across a range of ALK resistance mutations, including G1202R.
In our study, we collected real-world data for lorlatinib therapy in heavily pretreated Caucasian patients with a long follow-up time. We achieved a maturity of our data of 79.3% and 67.6% for the duration of treatment and OS for ALK-positive patients, and 57.1% and 42.9% for ROS1-positive patients.
Our findings showed a relevant efficacy of lorlatinib for ALK-positive NSCLC even after multiple lines of treatment matching the results of the pivotal trial, with a median DoT of 4.4 months,

Discussion
The optimal treatment sequence in the setting of ALK-positive NSCLC is of substantial clinical relevance. ALK TKIs are often particularly effective in these patients, thus changing the course of their disease, and their optimal sequence might affect patient outcomes to a considerable extent. As controlled clinical trials investigating various sequences of ALK TKIs across multiple treatment lines will presumably never become available, real-world data can add important information to the body of evidence. For instance, a retrospective analysis conducted across 23 Italian centers showed that sequential administration of crizotinib and a next-generation ALK TKI provided remarkable clinical benefits in a real-life population of 290 patients [29].
Lorlatinib appears to be an ideal choice in the later-line setting, given its uniquely broad activity against ALK resistance mutations. According to in-vitro data, most of the commonly observed secondary ALK mutations are inhibited by lorlatinib [13]. The pivotal clinical trial demonstrated rapid, deep and durable responses with lorlatinib after up to three prior ALK TKIs that had been administered with or without chemotherapy [21,28]. Furthermore, it has shown an improvement in global quality of life for the majority of patients treated with lorlatinib [30]. Hypercholesterolemia and hypertriglyceridemia represented the most common treatment-related AEs, and the antitumor activity of lorlatinib was seen across a range of ALK resistance mutations, including G1202R.
In our study, we collected real-world data for lorlatinib therapy in heavily pretreated Caucasian patients with a long follow-up time. We achieved a maturity of our data of 79.3% and 67.6% for the duration of treatment and OS for ALK-positive patients, and 57.1% and 42.9% for ROS1-positive patients.
Our findings showed a relevant efficacy of lorlatinib for ALK-positive NSCLC even after multiple lines of treatment matching the results of the pivotal trial, with a median DoT of 4.4 months, median OS from lorlatinib start of 10.2 months, median OS from Stage III/IV of 41.8 months and an ORR of 43.2%.
We looked further into the patients who received only one second generation ALK TKI (n = 20) and compared their OS to those who had more prior lines of second generation TKIs (n = 17). Although there has been a numeric difference in the ORR, with 50% versus 35% in favour of patients who have received only one second generation TKI, our analysis indicated that they are significantly similar. Comparing patients with different amounts of prior TKI lines in general showed no difference between groups. Lorlatinib even demonstrated activity when administered after brigatinib, which is of particular interest, as data on this sequence have been lacking to date. In the phase II pivotal study, most of the 139 patients included had received alectinib as their last ALK TKI treatment prior to lorlatinib (n = 62), whereas brigatinib had been administered prior to lorlatinib in only eight of the patients [17]. In our data, we identified 19 patients who had brigatinib as their last line of therapy. That patients showed an ORR of 32.6% and DoT of 3.5 months. Although there was no statistically significant difference between the groups, it should be pointed out, that patients who had alectinib prior to lorlatinib (n = 15) showed higher response rates and DoT (ORR 60.0%; DoT 8.5 m). This observation was still made when analyzing patients with only one second generation TKI prior to lorlatinib (prior alectinib ORR 66.7% and DoT 10,0m vs. prior brigatinib ORR 36.4% and DoT 3.5 m). These findings should be interpreted with care and need to be investigated in following studies. In conclusion, there was no significant difference in the DoT, OS and ORR regarding the line or sequence and lorlatinib may be administered in earlier as well as later lines of therapy and also after second generation TKIs.
A recently published study providing real-life data for lorlatinib in a later line setting for mainly Asian patients showed comparable results for the extracranial response rate (22%, n = 59) and median DoT (5.6 months) in ALK-positive patients. Differences can, however, be found concerning survival rates from diagnosis of advanced disease, as we reported a 2-year OS of 70.3% (comparing 95.8%, n = 76) and 5-year OS of 33.3% (compared to 79.4%). This can perhaps be partly explained by the difference in ethnicity and less heavily pretreated population (≥3 prior systemic lines in 90% of our ALK-patients vs. 67%) [31].
Moreover, our findings indicate notable activity of lorlatinib in individuals with ROS1 rearrangements. A higher proportion of these patients, relative to those with ALK-positive disease, were receiving ongoing treatment at the time of analysis, and their median OS was longer. However, the limited patient numbers preclude firm conclusions on this, and lorlatinib has not been approved for clinical use in patients with ROS1-positive disease, to date.
Lorlatinib treatment was well tolerated. The majority of events did not contribute significantly to the symptom burden, which is an important aspect in the later-line setting. Hypercholesterolemia and hypertriglyceridemia occurred as the most common events. Only one patient had to permanently discontinue treatment due to an AE. AE rates were lower than the rates reported in the clinical trial setting, which is due to the observational nature. Moreover, the retrospective nature of this analysis entails a series of general limitations.
Due to the lack of sufficient data to the exact time of radiologic disease progression, we have not been able to calculate the progression free survival (PFS). As a surrogate marker, we used DoT, including treatment beyond progression. As some of the patients had received lorlatinib in later lines, without many reasonable alternative treatment options, there might be a discrepancy between DoT and PFS.
The question of whether lorlatinib therapy prevented the development of brain metastases cannot be fully answered based on our observations here, because patients without known CNS lesions and without neurologic symptoms did not undergo brain imaging at baseline due to standard of care. Meaningful data for intracranial response (ORR 62.5%) was possible to collect only from a small number of patients who had confirmed brain metastases at initiation of lorlatinib and did not receive any cranial radiation.
In addition, we have not generated data on resistance mutations, as testing for these mutants is not routinely conducted, which is often due to the difficulty in obtaining tissue and the knowledge that next-generation ALK inhibitors appear to be active regardless of the presence of resistance mutations. Nevertheless, the jury is still out regarding biomarker-driven treatment decisions in ALK-positive lung cancer, and investigations in this area are ongoing. The ALK Master Protocol study is currently exploring the usefulness of mutation testing using liquid biopsies followed by mutation-specific ALK TKI treatments in patients with ALK-positive, stage IV NSCLC who have experienced progression on next-generation ALK inhibitors (NCT03737994).
Although the broad mutational coverage of lorlatinib predisposes it for later-line treatment, it is conceivable that this very feature provides pronounced clinical benefit in the first line setting. In fact, the pivotal trial that was conducted in both treatment-naïve and previously treated patients showed an ORR of 90% in the treatment-naive patient group [28]. In comparison, response rates were markedly lower in patients who received lorlatinib after one or more of the other ALK TKIs.
Overall, the findings presented here might contribute to filling a data gap with respect to later-line use of lorlatinib under real-world conditions. They suggest relevant clinical activity of this therapy even in massively pretreated patients, with good tolerability.

Materials and Methods
We conducted a retrospective, multicenter, real-world analysis on patients with ALKor ROS1-positive NSCLC who had undergone treatment with lorlatinib as second or later line therapy. Prior to lorlatinib, all ALK-positive patients had received at least one second generation ALK TKI. ROS1-positive patients had received at least one first or second generation TKI. The standard starting dose of lorlatinib was 100 mg/day.
Patients included were treated in 5 lung cancer centers in Austria as part of the Pfizer global compassionate use program between January 2016 and May 2020. All data were extracted from medical records. The primary outcome duration of treatment (DoT) was defined as the time from initiation of lorlatinib to last administered dose. Overall survival (OS) was defined as the time from first lorlatinib dose to date of death. In addition, we calculated the overall survival (OS2) defined as time from diagnosis of locally advanced or metastatic disease to death.
The best overall response was assessed by the treating investigators determined by radiologic complete response (CR), partial response (PR), stable disease (SD) and progressive disease (PD). Overall response rate (ORR) was defined as the sum of CR and PR, and disease control rate (DCR) was defined as the sum of CR, PR and SD compared to the overall number of patients.
Adverse events (AEs) were graded as per Common Terminology Criteria for Adverse Events (CTCAE) version 5.0.

Statistical Analysis
Median DoT and median OS as well as the number at risk were analyzed using the Kaplan-Meier-estimator and a confidence interval (CI) of 95%. Confidence interval for proportions like ORR was calculated using the Clopper-Pearson method.
The data cut-off was May 21, 2020. Patients were censored for the analysis if they were still alive for OS or on treatment for DoT. The median follow-up time was calculated from the time of first dose of lorlatinib to the data cut-off.
For a comparison of subgroups, the Log-rank test with a level of significance of 5% (chi square p = 0.05) was used. The first subgroup comparison in ALK-positive patients was done by separating the number of prior TKI lines, and the second subgroups are defined according to whether they received more than one second generation TKI. All statistical analysis where conducted using SPSS v.26.0 (IBM SPSS Statistics).

Conclusions
This retrospective analysis indicates that patients with advanced ALKor ROS1-positive NSCLC who have been previously treated with ALK-TKIs can be treated in later lines with lorlatinib for prolonged periods of time. Lorlatinib was well tolerated and was also active for the ROS1-positive population. Further evidence is required to confirm these observations. For the time being, the question remains as to whether to use third generation ALK inhibitors as later lines, rather than in the early treatment setting.
Author Contributions: M.J.H., H.F., O.I. and C.W. were responsible for study conception and drafting; H.F. was responsible for data analysis; R.W., E.B., E.W., M.S. were responsible for acquisition of data. All authors were responsible for revision of the manuscript. All authors have read and agreed to the published version of the manuscript.
Funding: This research received no external funding