Of the 54 compounds included in the analysis, all had dose modifications and discontinuations reported in some form. Overall, dose modifications due to adverse events (AEs) were high. Across all compounds, a median of 45.8% of patients reported at least one dose interruption due to AE(s), and 21.0% of patients reported at least one dose reduction due to AE(s) (Table 1). Forty-one compounds (76%) reported > 33% of patients with at least one dose interruption due to AE(s). Eleven compounds (20%) reported > 33% of patients with at least one dose reduction due to AE(s), with six compounds reporting > 50% of patients with at least one dose reduction due to AE(s). Treatment discontinuations due to AE(s) occurred in a median of 12.8% of patients across all compounds, ranging from 2 to 35% of patients.

Of the 54 compounds, 15 (28%) were approved at a dose equal to the identified MTD (Label Dose = MTD) and the remaining 39 (72%) were either approved at a dose lower than the MTD, or the MTD was not identified. Of these 39 compounds, 14 (26%) identified an MTD (Label Dose < MTD), 5 (9%) did not identify an MTD but were approved at the maximum studied dose (Label Dose = MSD), and 20 (37%) did not identify an MTD and were approved at a dose lower than the maximum studied dose (Label Dose < MSD). When comparing dose modifications by label dose classification, (Fig. 1A), compounds where the MTD was identified had a notably higher percentage of patients experiencing at least one dose interruption due to AE(s) (52.0% and 58.0% of patients for Label Dose = MTD and Label Dose < MTD, respectively) than compounds where the MTD was not identified (34.0% and 39.0% of patients for Label Dose = MSD and Label Dose < MSD) (Table 1). Of note, only one compound in the Label Dose = MTD group reported < 33% of patients undergoing at least one dose interruption, with most compounds in this group exceeding 50% of patients. Dose reductions due to AE(s) followed a similar trend when comparing by label dose classification, with a median percentage of patients experiencing at least one dose reduction of 28.0% and 24.0% in the Label Dose = MTD and Label Dose < MTD groups, vs. 5.0% and 10.5% in the Label Dose = MSD and Label Dose < MSD groups. Seven out of fifteen compounds in the Label Dose = MTD group (with reductions data available) reported > 33% of patients experiencing at least one dose reduction, compared to 4/36 compounds (with reductions data available) in the other groups.

Label dose classification of compounds with respect to MSD or MTD as described by dose modifications, reported as % of patients with drug discontinuation, or at least 1 dose interruption or dose reduction due to adverse events from the registrational trial (A). Relative Dose Intensity (RDI) reported from the registrational trial for each approved oncology small molecule (B). Percent of patients experiencing at least 1 Grade ≥ 3 TEAE while being administered the label dose of approved drug from the registrational trial (C). Label dose groups were compared for statistical significance by one-way ANOVA with Tukey’s post hoc HSD (* p < 0.05, ** p < 0.01, *** p < 0.001)

Interestingly, treatment discontinuations due to AE(s) were similar across groups when comparing by label dose classification, with a median of 13.0% and 19.5% of patients in the Label Dose = MTD and Label Dose < MTD groups, compared to 9.3% and 9.5% of patients in the Label Dose = MSD and Label Dose < MSD groups. Only 2 compounds (duvelisib and glasdegib) reported > 33% of patients discontinuing treatment due to AE(s), both in the Label Dose < MTD group. For duvelisib, the discontinuation rate was attributed primarily to diarrhea/colitis (10% of discontinuations), followed by rash (4%), pneumonia (3%), and pneumonitis (3%) [10]. For glasdegib, the discontinuation rate was attributed primarily to pneumonia (6%), QT interval prolongation (5%), sepsis (4%), and febrile neutropenia (4%) [10].

Overall, RDI was high, with a median RDI of ~ 98% reported for all compounds (Table 1). Most compounds had RDI > 90%, with only four compounds < 80%. No compounds reported a median RDI of < 70%. Of the 11 compounds within the Label Dose = MTD group (with median RDI data available), the median RDI was 91.9%, which was slightly lower than the median RDI of 95.0% for the Label Dose < MTD group, 100% for the Label Dose = MSD group, and 98.0% for the Label Dose < MSD group (Table 1). Overall, the Label Dose = MTD group skewed lower compared to the other groups, with 3 compounds reporting median RDI < 80% versus only 1 compound (which was Label Dose < MSD) reporting median RDI < 80% in all other groups (Fig. 1B). While the median RDI appears high for all groups, it should be noted that reporting of dose intensity was not standardized across compounds. Many compounds did not have a summary of dose intensity readily available (n = 12), and several other compounds reported dose intensity using methods other than mean or median RDI (n = 6). This bias should be taken into consideration in conjunction with reported dose modifications and discontinuations when interpreting overall compliance with the labeled dosing regimen.

Of the 54 compounds included in the analysis, AE data in terms of the total % of patients experiencing at least 1 Grade ≥ 3 TEAE at the approved dosing regimen was reported for all compounds, as was the incidence of specific Grade ≥ 3 TEAEs at the approved dosing regimen. Overall, Grade ≥ 3 TEAEs were high, with a median of 60.8% of patients experiencing a Grade ≥ 3 TEAE. Comparing the Label Dose = MTD vs. other groups, the Label Dose = MTD group reported a median of 63.7% of patients experiencing at least one Grade ≥ 3 TEAE, which was similar to the Label Dose < MTD (median 65.5% of patients) and Label Dose = MSD (median 68.1% of patients), but overall higher as compared to Label Dose < MSD group, with a median of 51.5% of patients, respectively (Fig. 1C).

To understand the driver of these high number of Grade ≥ 3 TEAEs, we looked at the Grade ≥ 3 TEAEs occurring at high incidence, defined as occurring in > 10% of patients at the approved dosing regimen. Of the compounds, 14/54 reported 0 high incidence Grade ≥ 3 TEAEs, 17/54 reported 1, and 23/54 reported ≥ 2 Overall, the high number of Grade ≥ 3 TEAEs appeared to be driven primarily by hematological events (Table 2), with 25/54 compounds reporting at least 1 hematological Grade ≥ 3 TEAE occurring in > 10% of patients. In most cases, high incidence hematological TEAEs occurred in multiples (for example a compound with > 10% of patients having Grade ≥ 3 thrombocytopenia may also have > 10% of patients having Grade ≥ 3 anemia), Other common high incidence Grade ≥ 3 TEAEs included gastrointestinal and cardiovascular events, occurring in > 5 compounds.

Comparing the relationship between percentage of patients experiencing at least 1 Grade ≥ 3 TEAE and treatment compliance, a moderate correlation between Grade ≥ 3 TEAEs and dose reductions, interruptions, and discontinuations was observed. However, no positive correlation between Grade ≥ 3 TEAEs and RDI was observed (Table 3).

Of the 54 compounds included in the analysis, a total of 7 compounds were issued a PMR for dose optimization at registration. Dose intensity, dose modifications and discontinuations due to AE(s), and incidence of Grade ≥ 3 TEAE were analyzed, grouped by compounds receiving a PMR for dose optimization (Fig. 2). Compounds issued a PMR for dose optimization trended lower in RDI and higher in dose modifications due to AE(s) and incidence of Grade ≥ 3 TEAEs than compounds not issued a PMR for dose optimization. Regarding dose modifications, a higher percentage of patients experiencing at least one dose modification due to AE(s) were reported for molecules issued a PMR in dose optimization vs. the latter (median 66.0 vs. 45.6% interruptions, and 45.0 vs. 20.1% reductions, respectively) (Table 4). The incidence of discontinuations due to AE(s) were similar among groups, where compounds receiving a PMR for dose optimization reported a median of 14.8% of patients discontinuing due to AE(s), whereas the latter reported a median of 11.1% of patients. The trend continued with dose intensity, as compounds issued a PMR for dose optimization reported a median RDI of 86.6%, notably lower than the median RDI of 98.1% reported in compounds without. Incidence of high grade TEAEs also followed the trend, wherein the compounds receiving a PMR for dose optimization a median of 77.0% of patients were reported as experiencing at least one Grade ≥ 3 TEAE vs. 55.2% of patients in the latter. Interestingly, none of the 7 compounds issued a PMR for dose optimization reported a positive exposure-response relationship in efficacy and only 4 out of 7 reported a positive exposure-response in safety events.

Compounds that received a PMR in dose optimization as described by dose modifications, reported as % of patients with drug discontinuation, or at least 1 dose interruption or dose reduction due to adverse events from the registrational trial (A), Relative Dose Intensity (RDI) reported from the registrational trial for each approved oncology small molecule (B), percent of patients experiencing at least 1 Grade ≥ 3 TEAE while being administered the label dose of approved drug from the registrational trial (C). PMR groups were compared for statistical significance by Welch two sample t-test (* p < 0.05, ** p < 0.01, *** p < 0.001)

Out of these 7 compounds with a PMR issued for dose optimization, 3 were in the Label Dose = MTD group (adagrasib, futibatinib, infigratinib) and 2 were in the Label Dose = MSD group (sotorasib, quizartinib) (Table 4). The 2 remaining compounds were ribociclib (Label Dose < MTD) and selinexor (Label Dose < MSD). The specific justification for issuing these 2 compounds (ribociclib and selinexor) a PMR in dose optimization has been discussed previously [11] and is also available in the FDA multidisciplinary review [10]. Briefly, for ribociclib, while the pivotal trial dose was below the identified MTD, a positive exposure-response relationship was identified with regard to QT interval prolongation, but a flat exposure-response relationship with efficacy. The dose modifications due to AE(s) for ribociclib were also higher than the all-compound median, at 44.6% of patients experiencing at least one reduction and 71.3% of patients experiencing at least one interruption. For selinexor, while the pivotal trial dose was below the MSD, the percentage of patients experiencing at least one dose reduction or interruption due to AE(s) were also higher than the all-compound median, reported at 43.1% and 69.3%. This compound reported a positive exposure-response relationship in safety events but did not report a positive exposure-response relationship in efficacy.

Two compounds that received a PMR for dose optimization, sotorasib and quizartinib, were exceptions to the RDI trend, as they both reported a median RDI of around 100%. Sotorasib was approved for KRAS G12C-mutated NSCLC at 960 mg QD, the MSD. With respect to dose modifications and discontinuations due to AE(s), the percentage of patients experiencing at least one reduction was 5%, interruption was 34%, and discontinuation was 9%, which were below the all-compound median reported in Table 4. Sotorasib ultimately received a PMR to test a lower dose level of 240 mg QD due to a relatively flat dose-exposure relationship between the clinically tested dose ranges of 180 mg and 960 mg QD, attributed to saturable absorption. Quizartinib was approved for newly diagnosed AML that is FLT3 internal tandem duplication positive at a starting dose of 35.4 mg QD in combination with induction cytarabine and anthracycline, cytarabine consolidation, and as a maintenance monotherapy following consolidation chemotherapy. During maintenance, quizartinib can be administered at 26.5 mg QD for the first 14 days if QTcF is ≤ 450ms and escalated to 53 mg QD on Day 15 if QTcF ≤ 450ms. Quizartinib reported a positive concentration relationship with QTcF prolongation and other safety events but did not show a positive exposure-response in efficacy with the 53 mg QD maintenance dose. Therefore, the FDA issued a PMR in dose optimization to test a lower maintenance dose of quizartinib [10]. Of note, the percentage of patients experiencing at least one AE-related quizartinib dose reduction was 18.9% and interruption was 34.0%, which are similar to the all-compound median reported in Table 4, though the discontinuation rate was slightly higher at 20.4%.