Real-world treatment patterns and subsequent treatment effectiveness following frontline brigatinib in the ALTA-1L trial.

1.
Introduction
The oncogenic anaplastic lymphoma kinase (ALK) gene rearrangement occurs at a frequency of approximately 3% to 7% of patients with non–small cell lung cancer (NSCLC) [1–3]. Tyrosine kinase inhibitors (TKIs) target the resultant rearranged ALK protein [1]. The TKIs alectinib, brigatinib, and lorlatinib are currently recommended for frontline treatment of advanced ALK+ NSCLC, with ceritinib and crizotinib suggested if the 3 recommended frontline therapies are not available [4,5]. However, there is a lack of consensus on the optimal sequencing of different ALK TKIs for frontline and second-line (2L) treatment of advanced ALK+ NSCLC [4,5].
Brigatinib is a next-generation ALK TKI that demonstrated superior efficacy over crizotinib for the frontline treatment of patients with advanced or metastatic ALK+ NSCLC in the global phase III ALTA-1L trial (clinicaltrials.gov identifier: NCT02737501) [6,7]. In ALTA-1L, patients randomized to the brigatinib arm had a significantly lower risk of disease progression or death compared with those in the crizotinib arm; median progression-free survival (PFS) by blinded independent review committee (BIRC) assessment was 24.0 months with brigatinib and 11.1 months with crizotinib (hazard ratio, 0.48; log-rank P < 0.0001) [7].
Data are lacking on treatment patterns after initial brigatinib therapy and clinical outcomes with subsequent lines of treatment. Information on subsequent anticancer therapies collected under the ALTA-1L trial protocol is limited. As such, real-world data around treatments and outcomes after frontline brigatinib are needed to inform clinical decision-making in the long-term care of patients with advanced ALK+ NSCLC. This post-trial chart-review study retrospectively evaluated treatment patterns and outcomes with subsequent treatments in patients treated with frontline brigatinib in ALTA-1L.

2.
Materials and methods
2.1.
Study design and patients
This study was a retrospective, longitudinal, noninterventional, multinational, multisite-based, chart review of patients who were previously enrolled in the brigatinib arm of the ALTA-1L trial and discontinued frontline brigatinib during or after completing the trial (Figure 1). All sites from ALTA-1L with a patient who received brigatinib were approached to participate in this study; patients were enrolled in the chart review only if their site agreed to study participation. Patient data from medical records were entered into the electronic case report form between 4 November 2022 (first patient in), to 18 October 2023 (database lock). Patients were followed from the date of their last dose of frontline brigatinib either before or after the end of ALTA-1L (index date) until the end of patient follow-up or death.

This study was conducted in accordance with the Declaration of Helsinki 9, International Society for Pharmacoepidemiology Good Pharmacoepidemiology Practices 10, and any applicable local regulations. Because the data were collected retrospectively, waivers for obtaining informed consent were requested from all participating sites during the regulatory review, when possible. If a waiver could not be obtained for a site, only data from patients who indicated their agreement to the collection of their personal data were collected. If required by local laws and regulations, eligible patients were approached to provide written informed consent for review of their medical records for study data collection.

2.2.
Study outcomes
Demographic and clinical characteristics and post-brigatinib treatment pattern and outcomes data were collected. The 2L real-world overall response rate (rwORR) was defined as the percentage of patients with best response of real-world complete response (rwCR) or real-world partial response (rwPR) during 2L treatment. When images were not available, tumor responses were determined using clinician notes. The real-world time to treatment discontinuation (rwTTD) of 2L treatment was defined as the time from 2L treatment initiation to discontinuation for any reason. This definition of rwTTD is consistent with that recommended for real-world studies by the Friends of Cancer Research [8]; the definition was used in a machine learning approach that predicted rwTTD [9] and in another analysis showing that rwTTD correlated positively with PFS in NSCLC [10]. Treatment discontinuation was defined as the date of the last drug administration during the same line of treatment or death, whichever occurred first. Patients were considered to have discontinued treatment if they had initiated a new line of treatment since last drug administration, died, or had no data for 120 days or more after last drug administration. The real-world PFS (rwPFS) on 2L treatment was defined as the time from initiation of the first subsequent treatment to real-world progressive disease (rwPD) or death, whichever occurred first. Patients were censored at the end of the treatment line or at the date of last contact. Real-world PFS2 (rwPFS2) was defined as the time from randomization to brigatinib in ALTA-1L to the first progressive disease (PD) documentation on the first subsequent (ie, 2L) systemic anticancer treatment (not including surgery or radiotherapy) or death in patients who initiated subsequent systemic anticancer therapy. Patients without documented rwPD or death were censored on the initiation date of a new line of treatment (ie, third-line treatment) or the last contact date. Overall survival (OS) was defined as the time from the date of randomization to brigatinib in ALTA-1L until death. Adverse events of special interest (AESIs) due to the frequency of their occurrence in ALK TKI clinical trials were assessed from the time of brigatinib discontinuation through initiation of 2L treatment until the end of follow-up or death; AESIs included alanine aminotransferase increase, anemia, aspartate aminotransferase increase, blood bilirubin increase, cognitive/mood effects, constipation, cough, creatine phosphokinase increase, diarrhea, gamma-glutamyl transferase increase, hypercholesterolemia, hypertension, hypertriglyceridemia, increased weight, lipase level increase, myalgia, nausea, peripheral edema, peripheral neuropathy, vision disorder, and vomiting.

2.3.
Statistical analyses
Patient characteristics and treatment patterns were summarized using descriptive statistics. Time-to-event analyses of 2L rwTTD, 2L rwPFS, rwPFS2, and OS were performed using the Kaplan-Meier (KM) method. The 95% confidence interval (CI) for the estimated median and KM estimates were calculated using the Brookmeyer and Crowley method. Tumor response and time-to-event outcomes were evaluated in the following subgroups of interest: patients treated with any 2L ALK TKI after frontline brigatinib; patients who received 2L lorlatinib, 2L alectinib, or 2L crizotinib; patients who received 2L non–ALK TKI therapy; and patients who were responders (ie, achieved best response of complete response [CR] or partial response [PR] by BIRC assessment) or nonresponders (ie, did not achieve best response of CR or PR by BIRC assessment) to frontline brigatinib among patients who received 2L ALK TKIs. Safety outcomes were evaluated in the following subgroups: any 2L ALK TKI, 2L lorlatinib, 2L alectinib, 2L crizotinib, and 2L non–ALK TKI therapy. Statistical analyses were performed using SAS version 9.4 (Cary, NC).

3.
Results
3.1.
Patients
As of 18 October 2023, 56 patients had been screened at 20 study sites in 10 countries (Austria, Denmark, France, Hong Kong, Italy, Netherlands, South Korea, Spain, Taiwan, and United Kingdom). Of these, 48 patients met inclusion criteria and comprised the enrolled population (Figure 2). The median age at the index date (last dose of frontline brigatinib) was 58 years, 54% of patients were female, 54% were Asian, and 63% had never smoked; 20 patients (42%) had brain metastases at the index date (Table 1). Histopathological classification at the time of ALTA-1L randomization was primarily adenocarcinoma (n = 44, 91.7%); other classifications included adenosquamous carcinoma (n = 1, 2.1%), squamous (n = 2, 4.2%), and other (n = 1, 2.1%). Median treatment duration of frontline brigatinib from the first dose in ALTA-1L to discontinuation before or after ALTA-1L study end (ie, index date) was 19.4 months (Q1, Q3: 5.0, 38.6) (Supplementary Table S1). The most common reason for discontinuation of frontline brigatinib was PD (71% [34/48]) (Supplementary Table S2). The median follow-up since the final dose of frontline brigatinib (index date) was 12.4 months (Q1, Q3: 3.7, 27.0).

3.2.
Subsequent treatment patterns
Forty (83%) of the 48 enrolled patients received subsequent systemic anticancer therapy (Table 2). The median time from the final dose of frontline brigatinib to start of 2L anticancer treatment was 4.0 days (Q1, Q3: 2.0, 21.5). Thirty of those 40 patients (75%) received another ALK TKI as their first subsequent therapy, most commonly lorlatinib (53% [16/30]), followed by alectinib (27% [8/30) and crizotinib (20% [6/30]). Eleven patients received 2L non–ALK TKI therapy, one of whom received it in combination with alectinib. Non–ALK TKI therapy included cisplatin or carboplatin combined with pemetrexed (n = 6), carboplatin + pemetrexed + pembrolizumab (n = 1), docetaxel (n = 1), carboplatin + gemcitabine (n = 1), and pimitespib (TAS-116; n = 1). Demographic and baseline characteristics of patients treated with 2L ALK TKIs or 2L non–ALK TKI therapy were consistent with the larger enrolled population (Table 1), although the proportion of patients with brain metastases at the index date was higher in patients receiving 2L ALK TKIs (53%) compared with those receiving 2L non–ALK TKI therapy (27%). Among the 30 patients who received 2L ALK TKIs, 23 patients (77%) had achieved a response (CR or PR) to frontline brigatinib and seven (23%) did not respond to brigatinib (Supplementary Table S3).

During their second line of therapy, 28% (11/40) of patients received radiotherapy; two patients (5%) had surgery, both for brain metastases.

3.3.
Tumor response to 2L treatment
Among patients treated with 2L ALK TKIs, the rwORR on 2L treatment was 33% and the real-world disease control rate (rwDCR) was 71% (Table 3). The rwORR was 31% with 2L lorlatinib, 57% with 2L alectinib, and 0% with 2L crizotinib. The rwDCR was 77% with 2L lorlatinib, 86% with 2L alectinib, and 25% with 2L crizotinib. None of the patients receiving 2L non–ALK TKI therapy had a response (rwORR: 0%), and two patients had stable disease (rwDCR: 22%).

For patients who had a response to frontline brigatinib and received subsequent treatment with an ALK TKI, the rwORR was 33% and the rwDCR was 67% (Supplementary Table S4). For patients who did not have a response to brigatinib, the rwORR was 33% and the rwDCR was 83% with 2L ALK-TKI treatment.

3.4.
rwTTD of 2L treatment
Among the 30 patients receiving 2L treatment with an ALK TKI, the median rwTTD of 2L treatment was 34.7 months (95% CI: 4.6–not reached [NR]) and 53% of patients (95% CI: 32%–70%) were still receiving the 2L ALK TKI at 2 years (Figure 3). The median 2L rwTTD was 37.2 months (95% CI: 6.0–NR) for lorlatinib, with 68% (95% CI: 35%–87%) of patients remaining on lorlatinib at 2 years, was NR (95% CI: 1.1–NR) for alectinib, with 50% (95% CI: 15%–77%) continuing alectinib at 2 years, and was 2.8 months (95% CI: 2.0–NR) for crizotinib, with 20% (95% CI: 1%–58%) continuing crizotinib at 2 years (Figure 3). Median rwTTD for patients treated with 2L non–ALK TKI therapy was 4.4 months (95% CI: 1.4–6.0), with no patients receiving 2L non–ALK TKI therapy at 2 years (Figure 3).

The median rwTTD of 2L ALK TKIs was 37.2 months (95% CI: 3.6–NR) in brigatinib responders and 11.6 months (95% CI: 2.8–NR) in nonresponders, with 57% (95% CI: 32%–76%) and 43% (95% CI: 10%–73%), respectively, remaining on 2L ALK TKI treatment at 2 years (Supplementary Figure S1).

3.5.
rwPFS for 2L treatment
The median rwPFS on any 2L ALK TKI treatment was 16.1 months (95% CI: 4.4–NR), 25.6 months (95% CI: 3.8–NR) for lorlatinib, 16.1 months (95% CI: 1.1–NR) for alectinib, and 2.4 months (95% CI: 2.0–NR) for crizotinib (Figure 4). The KM estimated 2-year rwPFS rate was 47% (95% CI: 26%–65%) for 2L ALK TKIs, 53% (95% CI: 24%–76%) for lorlatinib, 44% (95% CI: 10%–74%) for alectinib, and 40% (95% CI: 5%–75%) for crizotinib. For patients receiving 2L non–ALK TKI therapy, median rwPFS was 6.1 months (95% CI: 1.7–NR) and 2-year rwPFS was 0% (Figure 4).

In brigatinib responders versus nonresponders, the median rwPFS for 2L ALK TKIs was 25.6 months (95% CI: 3.8–NR; 2-year rwPFS: 51% [95% CI: 26%–71%]) versus 13.0 months (95% CI: 2.4–NR; 2-year rwPFS: 34% [95% CI: 5%–69%]) (Supplementary Figure S2).

3.6.
rwPFS2 for frontline brigatinib and 2L treatment
The median time to rwPD or death on 2L therapy (rwPFS2) from the date of randomization to frontline brigatinib in ALTA-1L was 51.6 months (95% CI: 25.9–NR) for patients who received 2L ALK TKIs, 74.7 months (25.9–NR) months for 2L lorlatinib, 47.2 months (95% CI: 12.6–NR) for 2L alectinib, and 24.5 months (95% CI: 3.5–NR) for 2L crizotinib (Supplementary Figure S3). KM estimated 2-year rwPFS2 was 78% (95% CI: 58%–90%) for 2L ALK TKIs, 87% (95% CI: 56%–96%) for 2L lorlatinib, 73% (95% CI: 28%–93%) for 2L alectinib, and 63% (95% CI: 14%–89%) for 2L crizotinib. Median rwPFS2 was 18.5 months (95% CI: 6.5–NR) and 2-year rwPFS2 was 42% (95% CI: 14%–69%) in patients who received 2L non–ALK TKI therapy after frontline brigatinib (Supplementary Figure S3).
Median rwPFS2 was 68.8 months (95% CI: 25.9–NR) in brigatinib responders who received 2L ALK TKIs (2-year rwPFS2: 81% [95% CI: 57%–93%]) and 51.6 months (95% CI: 3.5–NR) in brigatinib nonresponders treated with 2L ALK TKIs (2-year rwPFS2: 69% [95% CI: 21%–91%]) (Supplementary Figure S4).

3.7.
Overall survival
The median OS after randomization to brigatinib in ALTA-1L was 74.7 months (95% CI: 30.0–NR) for patients who received 2L ALK TKIs (14 [47%] of 30 patients died), 74.7 months (95% CI: 30.0–NR) for 2L lorlatinib (6 [38%] of 16 patients died), NR (95% CI: 13.8–NR) for 2L alectinib (4 [50%] of 8 patients died), and 22.9 months (95% CI: 8.2–NR) for 2L crizotinib (4 [67%] of 6 patients died). The estimated 3-year OS rate was 67% (95% CI: 47%–80%) for 2L ALK TKIs, 75% (95% CI: 46%–90%) for 2L lorlatinib, 75% (95% CI: 31%–93%) for 2L alectinib, and 33% (95% CI: 5%–68%) for 2L crizotinib. For patients receiving 2L non–ALK TKI therapy, median OS after brigatinib randomization was 21.3 months (95% CI: 10.3–36.3) and 3-year OS was 27% (7%–54%).
For brigatinib responders treated 2L with ALK TKIs, the median OS was NR (95% CI: 30.0 months – NR) with an estimated 3-year OS rate of 70% (95% CI: 47%–84%). Brigatinib nonresponders receiving 2L ALK TKIs had a median OS of 74.7 months (95% CI: 8.2–NR) and 3-year OS of 57% (95% CI: 17%–84%).

3.8.
Safety
Among the 30 patients who received 2L ALK TKIs, AESIs (AEs that occurred frequently in ALK TKI clinical trials) were reported in 14 patients (46.7%), with serious AESIs occurring in 1 patient (3.3%; diarrhea) (Table 4). Percentages of patients with AESIs were 56.3% for 2L lorlatinib (serious, 6.3%), 37.5% for 2L alectinib (serious, 0%), and 33.3% for 2L crizotinib (serious, 0%). The percentage of patients with AESIs was 9.1% in patients who received 2L non–ALK TKI therapy (serious, 0%).

4.
Discussion
To our knowledge, this is the first international study to examine subsequent treatment patterns and outcomes following discontinuation of frontline brigatinib in patients with advanced ALK+ NSCLC. Demographic and clinical characteristics of the subset of ALTA-1L patients who participated in this retrospective chart-review study were similar to those of the overall population enrolled in phase III ALTA-1L trial [6]. Greater than 80% of the patients were treated with subsequent systemic anticancer therapy following brigatinib discontinuation, usually an ALK TKI (75%). The most common ALK TKI administered after brigatinib was lorlatinib (53%; 16/30), followed by alectinib (27%; 8/30).
Real-world antitumor responses were observed with 2L ALK-TKI treatment post-brigatinib, but not with 2L non–ALK TKI therapy. The rwORR with 2L ALK TKIs was 33% and the rwDCR was 71%. Responses were observed with 2L lorlatinib (rwORR = 31%) and 2L alectinib (rwORR = 57%), but not with 2L crizotinib. In CROWN, a prospective phase III clinical trial, the investigator-reported overall response rate (ORR) with ALK TKIs as first subsequent therapy was 29% after frontline lorlatinib and 16% after frontline crizotinib [11]. The rwORR with lorlatinib post-brigatinib in the current study (31%) was consistent with that reported in a retrospective multicenter analysis that included 19 patients whose last line of prior therapy was brigatinib (ORR with subsequent lorlatinib was 33%), although these patients were more heavily pretreated (up to 4 prior lines) than those in the current study [12]. Other than case studies [13], there is a lack of previously published data on clinical response rates with alectinib after brigatinib.
rwTTD and rwPFS data from this study indicate that patients experienced prolonged clinical benefit from 2L treatment with an ALK TKI post-brigatinib. In contrast to the poor clinical outcomes observed with 2L non–ALK TKI therapy (median rwTTD: 4.4 months; median rwPFS: 6.1 months), the median rwTTD of 2L ALK TKI treatment was 34.7 months and median rwPFS from 2L ALK TKI treatment initiation was 16.1 months. The observation that rwTTD was nearly twice as long as rwPFS in patients treated with 2L ALK TKIs suggests that treating physicians maintained their patients on the same 2L ALK TKI for an extended period, even after disease progression on 2L treatment. Continuation of brigatinib treatment after disease progression was allowed at the investigator’s discretion in ALTA-1L [6]. Furthermore, continuation of ALK TKI treatment after progression has been shown to be effective [14,15]. In a retrospective analysis of a phase II trial, patients who continued lorlatinib beyond disease progression had a longer median OS than patients who discontinued lorlatinib post-progression [14]. A real-world retrospective analysis of patients treated with crizotinib or alectinib demonstrated continuation of treatment beyond disease progression was associated with slightly longer median OS [15].
In CROWN, median treatment duration of the first subsequent systemic anticancer therapy was 9.6 months (Q1, Q3: 2.9, 18.1) in 33 patients who had started a new treatment after lorlatinib and 13.3 months (Q1, Q3: 4.8, 21.2) in 103 patients receiving subsequent treatment post-crizotinib; the median duration of 2L treatment in patients who received a 2L ALK TKI was 9.6 months post-lorlatinib (n = 21) and 14.0 months post-crizotinib (n = 96) [16]. The median PFS on 2L treatment has not been reported for CROWN [16]. The reasons for the longer rwTTD of 2L ALK-TKI therapy observed post–frontline brigatinib in the current study compared with post-lorlatinib in CROWN are unclear. Both patient populations were small. Given the small number of patients evaluated in the current study, these results should be interpreted with caution. Furthermore, retrospective data obtained from real-world chart reviews differs from prospective data obtained from clinical trials, which are protocol driven.
In the current study, the median rwTTD of 2L treatment was 37.2 months for lorlatinib and NR for alectinib. Lorlatinib was associated with numerically longer median rwPFS (25.6 months) than alectinib (16.1 months) post-brigatinib, although the upper limits of the 95% CIs were NR with both drugs. Median rwTTD and rwPFS were numerically shorter with 2L crizotinib (2.8 months and 2.4 months, respectively). Substantially shorter median TTD values have been reported in previous real-world studies of subsequent treatment with lorlatinib post-brigatinib, most likely because the majority of patients in these studies had received multiple therapies prior to brigatinib [17,18]. In a retrospective study of 413 patients in the IQVIA longitudinal pharmacy claims database who received brigatinib between April 2017 and September 2020, 81% of patients had received ≥1 prior ALK TKI [17]. Approximately one third (57/167) of patients who discontinued brigatinib were treated with lorlatinib as their next line of therapy; for these patients, the median TTD of lorlatinib was 8.0 months (95% CI: 3.9–NR) [17]. Another retrospective study of patients treated with brigatinib as 2L or later therapy in the French Early-Access Program (August 2016–January 2019) reported that lorlatinib was given post-brigatinib for 68 of 92 of patients treated with one or more therapy after brigatinib; for these patients, the median duration of lorlatinib treatment was 5.3 months (95% CI: 3.6–7.6) [18]. Only case studies have described outcomes with alectinib in the post-brigatinib setting, and none have reported on alectinib post–frontline brigatinib [13].
This study examined real-world outcomes post–frontline brigatinib. Real-world outcomes after other ALK TKIs have been examined in similar studies. A retrospective study using hospital-based administrative claims data from Japan showed that the median TTD of ALK TKIs after failure of alectinib was 294 days (~9.7 months) between April 2021 and January 2023, when both lorlatinib and brigatinib were available for treatment in Japan [19]. In a retrospective review of medical records of patients who progressed on alectinib before receiving brigatinib, median TTD and PFS were 5.7 months and 4.4 months, respectively [20]. Another real-world study that evaluated outcomes in patients treated with 2L lorlatinib after frontline alectinib demonstrated a median TTD of 10.8 months [21]. Overall, TTD and PFS were shorter with ALK TKIs post-alectinib across these three studies compared with findings with ALK TKIs post-brigatinib in this study (median rwTTD, 34.7 months; median rwPFS, 16.1 months); however, the patient populations in these other studies were not always restricted to frontline alectinib, and patients could have received other treatments between alectinib and the subsequent ALK TKI [19,20]. Real-world data on ALK TKI treatment outcomes post-lorlatinib are currently lacking.
The median rwPFS2 (time to progression or death on 2L therapy from the start of frontline brigatinib) was 51.6 months for 2L ALK TKIs, with numerically longer median values in patients receiving 2L lorlatinib (74.7 months) compared with 2L alectinib (47.2 months), and shorter median values with 2L crizotinib (24.5 months). Estimated rates of OS at 3 years after randomization to brigatinib were more than twice as high with 2L ALK TKIs (67%) compared with 2L non–ALK TKI therapy (27%) and were similar with 2L lorlatinib (75%) and 2L alectinib (75%) but not crizotinib (33%). These data do not represent OS in the full ALTA-1L population, as patients who remained on brigatinib after ALTA-1L were not included in these analyses. OS data remained immature at the final analysis of ALTA-1L, with an estimated 4-year OS rate of 66% for all patients randomized to brigatinib [7]. Taken together with the rwPFS and rwTTD results, these data support the long-term clinical benefit of sequential treatment with second- or third-generation ALK TKIs and indicate that lorlatinib can provide substantial disease control post–frontline brigatinib.
Patients treated with 2L ALK TKIs with and without a response to prior frontline brigatinib in ALTA-1L had identical rwORRs with 2L ALK TKI treatment (33%), but responders had numerically longer median rwTTD of 2L ALK TKI treatment (37.2 vs 11.6 months), rwPFS (25.6 vs 13.0 months), and rwPFS2 (68.8 vs 51.6 months) than nonresponders; median OS was NR in responders and 74.7 months in nonresponders. While this suggests that patients who respond to brigatinib may have better clinical benefit with subsequent ALK TKI treatment compared with nonresponders, the generalizability of this finding is limited by the small number of brigatinib nonresponders (n = 7) in this analysis.
AESIs were reported in nearly half of patients who received 2L ALK TKI treatment post–frontline brigatinib. The incidence of AESIs was numerically higher with 2L lorlatinib (56.3%) than with other 2L ALK TKIs (alectinib, 37.5%; crizotinib, 33.3%) and 2L non–ALK TKI therapy (9.1%), and the only serious AESI (diarrhea) occurred in a patient receiving 2L lorlatinib.
Strengths of this study include the chart review approach, which allowed for the extraction of detailed clinical outcome data that reflect real-world standards of care. In addition, the duration of follow-up for subsequent treatment after frontline brigatinib is longer in this study than available in real-world claims databases, given that brigatinib was approved as frontline treatment in 2020 [22], and ALTA-1L began enrolling patients in 2016 [6].
There are also several limitations associated with the current analysis. First, patients from ALTA-1L were only included if their study site participated in this chart review, which may introduce site selection bias. The characteristics of the subset of patients from clinical sites participating in this study may have differed from the sites in the overall ALTA-1L trial and may not represent all sites in a specific country or the entire ALTA-1L population. Additionally, only patients who discontinued frontline brigatinib and went on to receive subsequent therapies were included, and these patients may have differed from the overall population enrolled in the brigatinib arm of ALTA-1L. The timing between discontinuation of frontline brigatinib and starting subsequent therapy was not protocol-specified and may have differed between patients; inclusion of this time may also introduce immortal time bias for outcomes assessed from frontline brigatinib randomization (ie, rwPFS2 and OS). There is a potential for systematic underreporting of information in patient medical charts. Given that this was a real-world study, clinician notes had to be relied on for tumor response assessments in cases where imaging was not available. The sample sizes of patients who enrolled in this study and received 2L treatment after brigatinib were relatively small. KM estimates of median times to events should be interpreted with caution, given the wide 95% CIs and small sample sizes. Additionally, KM estimates may be skewed due to censoring bias, immortal time bias, or gaps between progression and documentation in charts. As the goal was to describe the clinical journey after frontline brigatinib and the sample sizes were small, this study was designed to be a descriptive analysis and was not statistically powered to make comparisons between subgroups. Therefore, there was no matched control group and no adjustments for baseline differences between subgroups using stratification or methods such as propensity score matching or multivariable modeling; confounding cannot be ruled out, especially given the higher incidence of brain metastases between patients treated with 2L ALK TKIs versus 2L non–ALK TKI therapy.

5.
Conclusion
This is the first study evaluating long-term real-world outcomes following frontline brigatinib treatment. Most patients advanced to another systemic anticancer therapy after brigatinib discontinuation. The most common subsequent therapies were ALK TKIs, with lorlatinib being the most common targeted therapy. The treatment landscape of ALK+ NSCLC continues to evolve with expanded frontline treatment choices, such as increasing uptake of lorlatinib based on results from the CROWN study [23]. However, these findings suggest that brigatinib as frontline treatment followed by other ALK TKIs may be an effective treatment sequence in patients with ALK+ NSCLC. Patients treated with frontline brigatinib, followed by subsequent second- or third-generation ALK TKIs (including lorlatinib), experienced prolonged clinical benefits.

Supplementary Material

Supplemental Material