Retrospective analysis of arterial occlusive events in the PACE trial by an independent adjudication committee

Background The phase 2 PACE (Ponatinib Ph+ ALL and CML Evaluation) trial of ponatinib showed robust long-term benefit in relapsed Philadelphia chromosome-positive (Ph+) leukemia; arterial occlusive events (AOEs) occurred in ≥ 25% of patients based on investigator reporting. However, AOE rates vary depending on the definitions and reporting approach used. Methods To better understand clinically relevant AOEs with ponatinib, an independent cardiovascular adjudication committee reviewed 5-year AOE data from the PACE trial according to a charter-defined process and standardized event definitions. Results A total of 449 patients with chronic myeloid leukemia (CML) or Ph+ acute lymphoblastic leukemia (ALL) received ponatinib (median age 59 y; 47% female; 93% ≥ 2 prior tyrosine kinase inhibitors (TKIs); median follow-up, 37.3 months). The adjudicated AOE rate (17%) was lower than the non-adjudicated rate (i.e., rate before adjudication; 25%). The only adjudicated AOE in > 2% of patients was peripheral arterial occlusive disease (4%). Exposure-adjusted incidence of newly occurring adjudicated AOEs decreased over time. Patients with multiple baseline cardiovascular risk factors had higher adjudicated AOE rates than those without risk factors. Conclusions This independent adjudication study identified lower AOE rates than previously reported, suggesting earlier overestimation that may inaccurately reflect AOE risk with ponatinib. This trial was registered under ClinicalTrials.gov identifier NCT01207440 on September 23, 2010 (https://clinicaltrials.gov/ct2/show/NCT01207440).

of these responses with a 5-year overall survival rate of 73% for CP-CML [3]. However, arterial occlusive events (AOEs) were reported by investigators in 25% in the overall population (serious AOEs, 20%) and 31% in the CP-CML population (serious AOEs, 26%) in the 5-year follow-up [3]. The exposure-adjusted incidence of newly occurring AOEs decreased from year 1 (15.8 patients with events per 100 patient-years in the total population) to year 5 (3.9 per 100 patient-years) [3]. The incidence of AOEs associated with ponatinib use has varied widely in subsequent reports. Two retrospective studies have reported an absence or very low incidence (6%) of AOEs [4,5]. Other real-world studies have reported AOE rates ranging from 18 to 26% [6,7]. Multiple factors may contribute to variability in reported AOE rates, including differences in patient populations, as well as differences in the clinical definitions used to identify and categorize vascular occlusive events. One of the most important factors is the lack of a standardized approach for defining and capturing AOEs with BCR::ABL1 TKIs.
The AOE incidence rate reported for PACE was based on a list of approximately 400 Medical Dictionary for Regulatory Activities (MedDRA) preferred terms developed by the sponsor. However, differences in the preferred terms used to define AOEs led to variability in AOE incidence rates. Some preferred terms included in the AOE analysis of PACE are highly sensitive for identification of potential AOEs but may not themselves indicate the occurrence of arterial occlusions, frequently including symptoms or descriptions rather than events; these include chest pain, cold hands, dysarthria, and poor peripheral circulation. This approach to characterize AOEs based on adverse event terms results in broadly capturing non-specific symptoms that may be associated with AOE rather than true AOEs and may thus overestimate the incidence of clinically meaningful events.
A clear understanding of clinically relevant AOE risk is imperative when characterizing the benefit-risk profile of ponatinib. Patients with CP-CML who become resistant to a second-generation BCR::ABL1 TKI, either with or without a BCR::ABL1 gene mutation, generally experience low response rates and poor survival if treated with another second-generation TKI [8,9]. Importantly, ponatinib is the only currently available TKI effective in patients with the BCR::ABL1 T315I mutation [3]. Therefore, the potential for improved survival and duration of response on ponatinib may outweigh the risk of AOEs [8,9]. However, the lack of clear data regarding clinically meaningful AOEs has led to confusion about how to optimally use ponatinib to treat relapsed/refractory CML and Philadelphia chromosome-positive (Ph+) acute lymphoblastic leukemia (ALL) and, in some instances, avoidance in patients who could potentially benefit. To provide a more accurate characterization of AOE incidence with ponatinib, an independent adjudication committee of experts was convened to retrospectively adjudicate all AOE reports in the PACE trial in a standardized, rigorous manner.

PACE trial design
The phase 2 PACE trial (ClinicalTrials.gov identifier: NCT01207440) enrolled adults with CML or Ph+ ALL whose disease was resistant or intolerant to dasatinib or nilotinib, or who had the BCR::ABL1 T315I mutation regardless of prior TKI use [3]. All patients received ponatinib at a starting dose of 45 mg once daily (qd); dose reductions to 30 or 15 mg qd were applied per protocol (Table 1) to manage adverse events (AEs), or implemented proactively following recommendations from the sponsor in October 2013 in response to AOEs emerging as notable AEs. The trial has been completed; detailed methods are published [2,3].

Adjudication methods
All activities related to the adjudication of AOEs were conducted by ACI Clinical (Bala Cynwyd, PA), including the identification of an independent adjudication committee. ACI Clinical is a clinical research organization with expertise in Endpoint Adjudication and Data Monitoring Committees to support safety decisions around clinical development programs. ACI Clinical was contracted by the sponsor; adjudication activities were not part of the PACE trial.

Identification of AEs for adjudication
To ensure all relevant potential events were captured, the PACE AE dataset (449 patients with 12,224 AE records; extraction date: May 9, 2018) was searched using a comprehensive set of 604 preferred terms potentially relevant to AOEs that was developed by the sponsor (Table 2). This search strategy, which was more comprehensive than that used in initial analyses of the PACE trial, identified 181 patients and 455 AE records for adjudication (Fig. 1A). In addition, all patient deaths not attributable to disease progression by the clinical investigator were reviewed by the chair of the adjudication committee (described below) for identification of potential fatal AOEs. The adjudication committee identified 45 fatal events for review. In total, 202 patients and 490 events were submitted to the independent adjudication committee for review (Fig. 1A).
An individual case package containing all available clinical information (including medical history) was created for each event and provided to the adjudication committee members for their review. If a patient experienced more than 1 event within 48 h, these events were adjudicated as potentially representing a single clinical event, unless the case evidence suggested they were independent events. Individual events occurring > 48 h apart were adjudicated as independent events. All data were from the clinical trial database that was in SAS format and structured in conformance to CDISC SDTM format; no other source material was available.

Adjudication procedure
An adjudication committee of academic research clinicians who are highly experienced in adjudication activities in cardiovascular trials was appointed by ACI Clinical. The adjudication committee of 5 independent academic experts (3 cardiologists, 1 vascular medicine specialist, and 1 vascular neurologist) retrospectively adjudicated suspected cases of arterial occlusive events in the PACE study. The committee followed a predefined process outlined in the adjudication charter developed by ACI clinical. The charter defined the responsibilities of the adjudication committee and the adjudication endpoints using established definitions developed by the 2014 American College of Cardiology (ACC)/American Heart Association (AHA) guideline [10], and the definitions for cardiovascular and stroke outcomes developed by the Standardized Data Collection for Cardiovascular Trials Initiative (SCTI) and the US Food and Drug Administration [11,12]. All suspected AOEs identified in the PT search were assessed using the charter definitions (Table 3) for myocardial infarction; heart   failure if attributed to an AOE, which may include  coronary artery disease, arterial hypertension, cardiomyopathy, or myocardial infarction; hospitalization  for unstable angina; stroke and other cerebrovascular  events; and peripheral vascular disease. Any events meeting the criteria of these endpoints were considered adjudicated AOEs. Specific criteria were required (e.g., revascularization, change in cardiac biomarkers, diagnostic evidence as shown by computerized tomography scan, magnetic resonance imaging, etc.) to determine the presence of a clinical endpoint. The adjudication committee members were blind to ponatinib dose at the time of the event, whether dose modifications were made, and the investigator's opinion on AE causality.
During the adjudication process, the committee reviewed all potential AOEs, as well as any AEs identified in a Cardiac Failure Standard MedDRA Query (SMQ), to determine whether any heart failure events were AOEs. Two members of the adjudication committee independently evaluated whether an individual case met the prespecified event definitions (Fig. 1B). If agreement between 2 members was not reached for cases of AOEs or heart failure, the case was reviewed by a third cardiologist adjudication committee member; if agreement was not reached with 3 votes, the case was reviewed at a panel meeting. If agreement was not reached for cases of stroke, deep vein thrombosis, pulmonary embolism, and peripheral vascular disease, the case was discussed at a panel meeting with the appropriate neurologist and/ or vascular specialist member(s). All fatal events were decided by consensus of adjudicators.
Events that met one of the charter-defined endpoint definitions were further categorized depending on the event type (e.g., myocardial infarction, peripheral arterial occlusive disease, deep vein thrombosis, etc.). Nonadjudicated AOEs that were recorded as symptoms (e.g., Table 1 Dose reduction recommendations (as of 2013)

Dose reduction recommendations
In October 2013, the following specific recommendations were formulated after discussions with the US FDA on evolving observations of arterial occlusive events in patients treated with ponatinib: All chronic phase chronic myeloid leukemia (CP-CML) patients on study who already had achieved major cytogenetic response (MCyR) should have had their dose reduced to 15 mg daily, unless, in the judgment of the investigator, the benefit/risk analysis, taking into account the patient's disease characteristics, BCR::ABL mutation status, and the patient's cardiovascular risk justified treatment with a higher dose All CP-CML patients on study who had not yet achieved MCyR should have had their dose reduced to 30 mg daily, unless, in the judgment of the investigator, the benefit/risk analysis, taking into account the patient's disease characteristics, BCR::ABL mutation status, and the patient's cardiovascular risk justified treatment with a higher dose All acute phase chronic myeloid leukemia (AP-CML), blast phase chronic myeloid leukemia (BP-CML), and Ph+ acute lymphoblastic leukemia (ALL) patients on study should have had their dose reduced to 30 mg daily, unless, in the judgment of the investigator, the benefit/risk analysis, taking into account the patient's disease characteristics, BCR::ABL mutation status, and the patient's cardiovascular risk justified treatment with a higher dose All patients who lost response at a lower dose may have their dose escalated (up to a maximum of 45 mg daily) as long as the dose was not lowered as a result of an adverse event (AE)   "non-cardiac chest pain" or "claudication") with a low severity level and no accompanying changes in medication or hospitalization were adjudicated to not be AOEs unless they had an anatomic diagnosis provided (e.g., "severe superficial femoral artery stenosis"). If the term "infarction" was provided for stroke events, the adjudicators categorized the event as ischemic stroke. Revascularization was not always clearly reported by investigators.

Statistics
Exposure-adjusted AOE rates were calculated as: (number of first events in interval)/(total exposure for interval in patient-years) × 100. The relative risk of serious AOEs was analyzed by baseline risk category in patients from the safety population for whom all baseline risk categories were available. Risk categories included commonly recognized cardiovascular risk factors for which data were collected (arterial hypertension, hypercholesterolemia, diabetes mellitus, and obesity), and history of heart disease (non-ischemic or ischemic).

Patient disposition and baseline characteristics
Patient disposition and baseline characteristics in the PACE trial have been published [2,3]. A total of 449 patients, including 270 CP-CML patients, 85 accelerated-phase (AP) CML patients, 62 blast-phase (BP) CML patients, and 32 Ph+ ALL patients, were enrolled  . c Per the charter, panel meetings were convened to discuss events for which a decision was not reached via independent voting. The quorum for panel meeting attendance was dependent on the type of event(s) to be discussed (i.e., cardiologist, neurologist, or vascular specialist) Table 3 Adjudication committee prespecified definitions of events

Cardiovascular (CV) death
The cause of death will be determined by the principal condition that caused the death, not the immediate mode of death. Members of the adjudication committee will review all available information and use their clinical expertise to adjudicate the cause of death CV death includes death resulting from an acute myocardial infarction (MI), sudden cardiac death, death due to heart failure (HF), death due to stroke, death due to CV procedures, death due to CV hemorrhage, death due to pulmonary embolism, and death due to other CV causes Death associated with acute myocardial infarction Refers to a death by any CV mechanism (e.g., arrhythmia, sudden death, heart failure, stroke, pulmonary embolus, peripheral arterial disease) ≤ 30 days after a MI related to the immediate consequences of the MI, such as progressive heart failure or recalcitrant arrhythmia. Acute MI should be verified to the extent possible by the diagnostic criteria outlined for acute MI (see below) or by autopsy findings showing recent MI or recent coronary thrombosis Death resulting from a procedure to treat a MI (percutaneous coronary intervention (PCI), coronary artery bypass graft surgery (CABG), or to treat a complication resulting from MI, should also be considered death due to acute MI Death resulting from an elective coronary procedure to treat myocardial ischemia (i.e., chronic stable angina) or death due to a MI that occurs as a direct consequence of a CV investigation/procedure/operation should be considered as a death due to a CV procedure Sudden cardiac death Sudden cardiac death refers to death that occurs unexpectedly, not following an acute MI (as defined above) and includes the following deaths: Witnessed and occurring without new or worsening symptoms Witnessed within 60 min of the onset of new or worsening cardiac symptoms, unless the symptoms suggest acute MI Witnessed and attributed to an identified arrhythmia (e.g., captured on an electrocardiographic (ECG) recording or witnessed on a monitor, or unwitnessed but found on implantable cardioverter-defibrillator review) After unsuccessful resuscitation from cardiac arrest (e.g., implantable cardioverter-defibrillator [ICD] unresponsive sudden cardiac death, pulseless electrical activity arrest) After successful resuscitation from cardiac arrest and without identification of a specific cardiac or non-cardiac etiology Unwitnessed death in a subject seen alive and clinically stable ≤ 24 h prior to being found dead without any evidence supporting a specific non-CV cause of death (information regarding the patient's clinical status preceding death should be provided, if available) Note: Unless additional information suggests an alternate specific cause of death (e.g., Death due to other CV causes), if a patient is seen alive ≤ 24 h of being found dead, sudden cardiac death should be recorded. For patients who were not observed alive within 24 h of death, undetermined cause of death should be recorded (e.g., a subject found dead in bed, but who had not been seen by family for several days) Note: Successful resuscitation without death should be captured as a resuscitated sudden cardiac death in the non-fatal voting flow Death due to HF Refers to death associated with clinically worsening symptoms and/or signs of HF regardless of etiology. Deaths due to HF can have various etiologies, including single or recurrent MIs, ischemic or non-ischemic cardiomyopathy, hypertension, or valvular disease Note: Due to the pro-thrombotic nature of the subject population, a thrombo-embolic option is included during voting. See rules in the non-fatal heart failure definition Death due to stroke Refers to death within 30 days that is either a direct consequence of the stroke or a complication of the stroke. Acute stroke should be verified to the extent possible by the diagnostic criteria outlined for stroke Death due to CV procedures Refers to death caused by the immediate complications of a cardiac procedure not in the context of treatment for acute MI Death due to CV hemorrhage Refers to death related to hemorrhage such as a non-stroke intracranial hemorrhage, non-procedural or nontraumatic vascular rupture (e.g., aortic aneurysm), or hemorrhage causing cardiac tamponade Death due to other CV causes Refers to a CV death not included in the above categories but with a specific, known cause (e.g., pulmonary embolism or peripheral vascular disease (venous or arterial disease) Non-CV death Non-CV death is defined as any death with a specific cause that is not thought to be of CV nature. Adjudication committee members will be asked to indicate the most likely cause of non-cardiovascular death on their voting form

Events Definitions
Examples of non-CV death are: pulmonary causes, renal causes, gastrointestinal causes, hepatobiliary causes, pancreatic causes, infection (including sepsis), inflammatory (e.g., systemic inflammatory response syndrome (SIRS))/immune (including autoimmune)(may include anaphylaxis from environmental (e.g., food allergies), hemorrhage that is neither cardiovascular bleeding or stroke, non-CV procedure or surgery, trauma, suicide, non-prescription drug reaction or overdose, prescription drug reaction or overdose (many include anaphylaxis), neurological (non-cardiovascular), malignancy (i.e., new malignancy, worsening of prior malignancy) or other (should be specified) Undetermined cause of death Undetermined cause of death refers to a death not attributable to one of the above categories. Inability to classify the cause of death may be due to lack of information (e.g., the only available information is "patient died") or when there is insufficient supporting information or detail to assign the cause of death. In general, most deaths should be classifiable as CV or non-CV, and the use of this category of death, therefore, should be discouraged and should apply to few patients in well-run clinical trials   Imaging demonstration of new loss of viable myocardium or new regional wall motion abnormality MI associated with stent thrombosis or stent restenosis as documented by angiography or at autopsy will also be captured as subtypes 4b and 4c  Imaging evidence of new loss of viable myocardium or new regional wall motion abnormality Heart failure event A heart failure event includes hospitalization for heart failure and may include any urgent outpatient visits for heart failure. The date of this event will be the day of hospitalization of the patient (including any overnight stay at the emergency room or chest pain unit) or the day of visit to the urgent outpatient center. Due to the prothrombotic nature of the subject population, a thrombo-embolic option is included during voting The following rules may be applied to indicate if heart failure is attributed to an AOE/VTE: Heart failure may be attributed to an AOE/VTE if related to coronary artery disease, hypertension, cardiomyopathy or myocardial infarction The relationship of heart failure to an AOE/VTE may be excluded if the underlying cause of heart failure is heart valve disorders, congenital heart disorders or arrhythmias Heart failure requiring hospitalization Heart failure hospitalization is defined as an event that meets all the following criteria: Patient is admitted to the hospital with a primary diagnosis of HF  Mechanical circulatory support (e.g., intra-aortic balloon pump, ventricular assist device, extracorporeal membrane oxygenation, total artificial heart) Mechanical fluid removal (e.g., dialysis, ultrafiltration, hemofiltration) Urgent heart failure visit An urgent heart failure visit is defined as an event that meets all the following criteria: The patient has an urgent, unscheduled office/practice or emergency department visit for a primary diagnosis of heart failure, but not meeting the criteria for a heart failure hospitalization All signs/symptoms for heart failure hospitalization (i.e., symptoms, physical examination findings/lab evidence of new or worsening HF as indicated under definition for Heart Failure Hospitalization) must be met The patient receives initiation or intensification of treatment specifically for heart failure, as detailed in the heart failure hospitalization section with the exception of oral diuretic therapy (which will not be sufficient) Hospitalization for unstable angina The date of this event will be the day of hospitalization of the patient including any overnight stay at an emergency room or chest pain unit Hospitalization for unstable angina is defined as an event that meets all the following criteria: Negative cardiac biomarkers and no evidence of acute MI Ischemic discomfort (angina or other symptoms thought to be equivalent) ≥ 10 min in duration occurring at rest or in an accelerating pattern with frequent episodes associated with progressively decreased exercise capacity

Events Definitions
Unscheduled hospitalization within 24 h of the most recent symptoms. Hospitalization is defined as an admission to an inpatient unit or a visit to an emergency department that results in at least a 24 h stay (or a change in calendar date if the hospital admission or discharge times are not available) At least one of the following:

Events Definitions
No other readily identifiable non-stroke cause for the clinical presentation (e.g., brain tumor, trauma, infection, hypoglycemia, peripheral lesion) Confirmation of the diagnosis by at least one of the following:** Neurology or neurosurgical specialist Brain imaging procedure (at least one of the following): CT scan

MRI scan
Cerebral vessel angiography Lumbar puncture (i.e., spinal fluid analysis diagnostic of subarachnoid hemorrhage) **If a stroke is reported but evidence of confirmation of the diagnosis by the methods outlined above is absent, the event will be discussed at a full EAC meeting. In such cases, the event may be adjudicated as a stroke on the basis of the clinical presentation alone, but full EAC consensus will be mandatory Undetermined stroke is defined as an acute episode of focal or global neurological dysfunction caused by presumed brain, spinal cord, as a result of hemorrhage or infarction but with insufficient information to allow categorization as #1 and #2 above Note: Given the scope of this study, stroke disability will not be measured. TIA definition was intentionally left out for this study; suspected TIA events will be identified for adjudication in order to rule out stroke

Venous thrombosis Superficial vein thrombosis
Superficial vein thrombosis (SVT) refers to a blood clot in one of the superficial veins near the surface of the body. There is usually an inflammatory reaction around the vein and may present with as a painful induration with erythema. An SVT can lead to a serious complication such as a higher risk for pulmonary embolism Superficial vein thrombosis could be documented by one of the following: Clinical symptoms (such as warmth, edema, 'cord-like' palpable mass, erythema, pain)

Duplex ultrasound
Deep vein thrombosis Deep vein thrombosis (DVT) refers to a blood clot in one of the deep veins (to include distal and proximal DVT). It may occur anywhere in the body but is most common in the extremities, a clot blocks blood circulation through these veins, which carry blood back to the heart. This commonly causes pain and swelling distal to the thrombus.  Peripheral vascular disease refers to a blood circulation disorder outside of the heart and brain that causes the blood vessels to block, narrow or spasm. PVD can be either in veins or arteries. Physical symptoms may include weak pulses, wounds/ulcers that won't heal, thin or pale skin PVD could be documented by one of the following:

Doppler ultrasound
Ankle-brachial index Angiography Magnetic resonance angiography Computerized tomography angiography Members will be asked to choose if this is a venous or arterial occlusive event Revascularization procedures For fatal and non-fatal cardiovascular endpoint events, members must also indicate if the event is associated with a revascularization procedure (PCI, CABG or PVI)

Percutaneous coronary intervention (PCI)
Defined as the placement of an angioplasty guidewire, balloon, or other device (e.g., stent, atherectomy, brachytherapy or thrombectomy catheter) into a native coronary artery or CABG for the purpose of mechanical coronary revascularization. The assessment of coronary lesion severity by intravascular ultrasonography, coronary flow reserve, or fractional flow reserve is not considered a PCI procedure

Coronary artery bypass graft (CABG)
Defined as a procedure performed to bypass partially or completely occluded coronary arteries with veins and/ or arteries harvested from elsewhere in the body, thereby improving the blood supply to the coronary circulation supplying the myocardium

Peripheral vascular intervention (PVI)
Peripheral vascular intervention is a catheter-based or open surgical procedure designed to improve arterial or venous blood flow or otherwise modify or revise vascular conduits. Procedures may include, but are not limited to percutaneous transluminal balloon angioplasty, stent placement, thrombectomy, embolectomy, atherectomy, dissection repair, aneurysm exclusion, treatment of dialysis conduits, placement of various devices, intravascular thrombolysis or other pharmacotherapies, and open surgical bypass or revision between September 2010 and October 2011. Baseline characteristics are summarized in Table 4. Among all 449 patients, the median age was 59 years and 53% of patients were male. Most (93%) patients had received 2 or more prior TKIs. At baseline, 53% of patients had arterial hypertension, 49% had hypercholesterolemia, and 24% had BMI ≥ 30 kg/m 2 . Forty-three percent of patients had a baseline history of non-ischemic cardiac disease, and 23% had a history of ischemic cardiovascular disease. Safety data reviewed by the adjudication committee reflect data collected as of February 6, 2017, with median follow-up of 37.3 months for all patients and 56.8 months (range 0.1-73.1 months) for CP-CML patients.
The exposure-adjusted incidence of adjudicated AOEs (8.9 patients with events per 100 patient-years) and serious AOEs (8.4 patients with events per 100 patientyears) was lower than the exposure-adjusted incidence of non-adjudicated AOEs (11.3 and 9.2 per 100 patientyears, respectively). The exposure-adjusted incidence of newly occurring AOEs decreased over time (Fig. 2B). The median time to onset of the first adjudicated AOE was 14.1 months (range: 0.1 to 49.5; Table 8).

Resolution of AOEs, dose modifications, and discontinuations
Among the 78 patients with an adjudicated AOE, events resolved in 51 patients. Among 43 patients with just one AOE, 74% (32/43) had resolution of the event; 35 patients had multiple AOEs recorded, with 54% (19/35) patients having resolution of all the events. Most patients continued ponatinib after the AOE, including 36 patients (46%) who continued ponatinib without dose modification and 27 patients (35%) who had their doses reduced and/or interrupted after the event (Table 9). Seven patients (9%) discontinued ponatinib due to an adjudicated AOE. Rates of dose modifications following AOEs are summarized in Table 9.

Risk factor analysis
The most common baseline risk factors in patients who developed an AOE were arterial hypertension and hypercholesterolemia (Table 10). Patients with adjudicated AOEs also had higher rates of concomitant use of antihypertensive medications, platelet aggregation inhibitor medications, and anti-diabetic agents compared with patients who did not have AOEs (Table 11).
The incidence of adjudicated AOEs by number of baseline risk factors (including arterial hypertension, hypercholesterolemia, obesity, diabetes mellitus, non-ischemic cardiac disease, and ischemic disease) is shown in Fig. 2C. The rate of adjudicated AOEs was 13% (24/189) among patients with 1-2 risk factors, and 29% (52/180) among patients with 3 or more risk factors. Of the 80 patients without any risk factors at baseline, only 2 (3%) had an AOE.

Fatal AOEs
Separate adjudication of deaths revealed that 11 adjudicated AOEs were associated with death. These included 2 cases of cardiac arrest and 1 each of the following: bradycardic arrest, cardiac failure, intracranial hemorrhage, worsening of congestive heart failure, superior mesenteric artery occlusion, hemorrhagic cerebral infarction, congestive heart failure, ischemic stroke, and acute anterior myocardial infarction. Nine of the 11 patients with AOEs associated with death had a history of cardiovascular events and/or cardiovascular risk factors recorded at baseline (Table 12). The long-term survival of patients with adjudicated AOEs was similar to survival of patients without AOEs (Fig. 2D).

Discussion
In this study, adjudication of AOEs by an independent committee of experts allowed for a clinically meaningful description of AOEs associated with ponatinib, which can help to inform health care providers and patients of safety risks in an accurate and objective manner. The search that identified potential AOEs for adjudication was broader (based on 604 MedDRA terms related to vascular ischemia or thrombosis) than that initially used to calculate non-adjudicated AOE rates in the PACE trial (400 MedDRA terms) [3]. Based on 5-year followup of the PACE trial, the adjudicated AOE rate (17%) was lower than the non-adjudicated AOE rate (25%) [3]. Although the majority of adjudicated AOEs were serious, 81% of patients with AOEs continued on ponatinib (35% with dose modifications), the benefit of the drug was felt to outweigh the risk of the AOEs. Although vascular occlusive events were rarely reported during the initial development of second-generation BCR::ABL1 TKIs, a meta-analysis found that these events occurred in 5.9% of patients with CML treated with these agents, including bosutinib, dasatinib, nilotinib, and ponatinib [13]. In another review of prospective trials of patients treated with TKIs, including imatinib, nilotinib, dasatinib, and ponatinib, overall incidence of CV events was 45% (range, 41-63%) [14]. Accordingly, a high level of vigilance is indicated to recognize this potential complication of TKI therapy.
Notably, although concern existed around the potential for increasing AOE rates with long-term dosing, as seen with AEs related to other TKIs [15][16][17][18] the exposureadjusted incidence of newly occurring adjudicated AOEs decreased over time on ponatinib, suggesting that the toxicity of ponatinib may not increase with longer treatment duration.
Patients with adjudicated AOEs were more likely to have multiple baseline cardiovascular risk factors (e.g., ischemic cardiac disease, arterial hypertension,   [3]. C Incidence of AOEs (adjudicated) by number of baseline risk factors (all patients). Risk factors included arterial hypertension, hypercholesterolemia, obesity, diabetes mellitus, non-ischemic cardiac disease, and ischemic disease. D Overall survival (OS) in chronic-phase chronic myeloid leukemia (CP-CML) patients with and without AOEs hypercholesterolemia, and diabetes mellitus), and only 2 patients had an adjudicated AOE without any cardiovascular risk factors. These observations align with those of previous studies [6,19]. It is important to identify and manage cardiovascular risk factors before and during therapy with ponatinib or other TKIs [20][21][22]. In PACE, 80% of CP-CML patients were resistant to dasatinib or nilotinib, and 24% had the BCR::ABL1 T315I resistance mutation [3]. Among CP-CML patients, estimated 5-year PFS and OS rates were 53% and 73%, respectively [3]. Data for overall survival in patients with and without adjudicated AOEs suggest that the risk of AOE-related death did not substantially impact survival, with diseaserelated death being the main driver of the OS curve. This underscores the need for providers to fully understand the therapeutic profile of ponatinib and consider its use when the potential benefits outweigh the risks for a given patient.
This study reinforces the importance of proper assessment of cardiovascular AEs to ensure accurate estimation of cardiovascular risk. The conventional processes of AE reporting and causality assessment may need to be re-assessed to avoid pitfalls associated with over-or under-reporting of AOEs, both of which may adversely affect patient care [23,24]. Formal adjudication of events is a mainstay for development programs in other therapeutic areas such as diabetes mellitus [25,26] and cardiology. A better understanding of the AOE risk associated with TKI therapy is a prime example of where formal adjudication is critical because accurate knowledge of risks is crucial before prescribing any TKI. The potential benefits of effective BCR::ABL1 TKI treatment, even with accompanying AEs, may outweigh the potential risks of progression-related mortality in patients with CP-CML and Ph+ ALL receiving second-or thirdline therapy. This is particularly true for patients such as those with the BCR::ABL1 T315I mutation who may have limited treatment options [27]. Understanding the true incidence of the most significant events is a central element in properly assessing the benefit-risk ratio of an intervention. All later-generation TKIs are  associated with risk of cardiovascular AEs [28], and the results of the formal adjudication process suggest the risk of these events with ponatinib may not be dissimilar to the event rates seen with some second-generation BCR::ABL1 TKIs [16][17][18].
A noteworthy finding in our analysis is that the exposure-adjusted incidence of newly occurring adjudicated AOEs decreased over time on ponatinib. These results are reassuring that the rate of new AOEs may not increase with longer duration of ponatinib treatment. Furthermore, patients with positively adjudicated AOEs were much more likely to have baseline    (27) 88 (23) 19 (26) cardiovascular risk factors (e.g., arterial hypertension, hypercholesterolemia, diabetes mellitus) or established cardiovascular disease; of those patients without any cardiovascular risk factors only 2 had a subsequent AOE. These results may provide clinical guidance with respect to the approach to use of ponatinib in patients at risk for an AOE. The ongoing phase 2 OPTIC trial (ClinicalTrials.gov Identifier: NCT02467270) is using a response-based dose reduction protocol approach to evaluate the optimal ponatinib dosing regimen for maximizing efficacy while mitigating toxicity. Results show that higher doses of ponatinib were associated with increased incidence of AOEs, with exposure-adjusted treatment-emergent AOE rates of 5.6%, 3.6%, and 2.1% for the 45-mg, 30-mg, and 15-mg cohorts, respectively [29]. However, the benefit differential was considerably larger with a starting dose of 45 mg, which was associated with a 26.3 percentage-point improvement in the response rate compared with a 15-mg starting dose (51.6% vs. 25.3%) [29]. Overall, the study indicated the best risk/benefit ratio when the 45-mg starting dose was reduced to 15 mg upon achievement of response (BCR::ABL1 IS transcript levels ≤ 1%) [29]. This retrospective study has strengths and limitations. The adjudication methodology provided a comprehensive and objective approach for characterizing AOE risk. A limitation is that only data from the clinical trial database were available. Prospective implementation of this strategy, as is being done in 2 ongoing trials, OPTIC and Ponatinib-3001 (NCT03589326), will overcome this challenge and add further value to the methodology and strength to the conclusions. In OPTIC, an independent cardiovascular endpoint adjudication committee is reviewing AOEs as they are reported using source documentation including cardiovascular workup (e.g., echocardiograms, electrocardiograms, biomarkers), hospitalization records, and any cardiovascular examinations performed.