A comparison of 12- and 19-week CHOP protocols using non-randomized, contemporaneous controls

2.1 Patient selection

From December 2015 to December 2017 pet owners at our institution were offered CHOP19 or CHOP12 for dogs diagnosed with multicentric lymphoma. The decision as to which treatment protocol to pursue was left to the owners, with guidance from the attending veterinarian regarding associated costs and risks of the protocols.

This study was prospectively conceived and while data were collated and analysed retrospectively, the criteria for inclusion/exclusion were established a priori and the patients were treated contemporaneously. Patients were included in study analysis if they were diagnosed with multicentric nodal lymphoma through cytology or histopathology of peripheral lymph nodes, with further characterization through polymerase chain reaction for antigen receptor rearrangements (PARR) or flow cytometry if elected by the owner. Patients with lymphoma that were primarily extranodal or with debilitating systemic illness prior to lymphoma diagnosis (eg, severe renal or hepatic compromise) were excluded. Other inclusion criteria required that the patient had not received previous treatment with chemotherapy or corticosteroids. Information regarding response to therapy and patient outcomes were collected from our institution's medical records system and records from referral partners.

2.2 Data collection

Information collected from medical records included breed, gender, neuter status, age at diagnosis, weight at diagnosis, date diagnosed, date treatment initiated, diagnosis method, complete blood count and serum biochemistry parameters, and chest radiograph and abdominal ultrasound abnormalities, if performed. Data regarding induction protocols, date of relapse, rescue protocols and date and reason of euthanasia or death were recorded.

Diagnosis of intermediate or large cell lymphoma was based on interpretation of cytological or histopathological assessment by a board-certified pathologist. Flow cytometry or histopathology was pursued on patients with cytologically intermediate cell lymphoma that was concerning for an indolent lymphoma prior to delivery of a CHOP-based protocol. Patients diagnosed with or suspected of having small cell lymphoma were excluded, as were patients suspected to have lymphocytic or lymphoblastic leukaemia. Immunophenotype and method of determination were recorded if available. Methods of immunophenotyping included PARR, flow cytometry, immunohistochemistry and immunocytochemistry.

Abnormalities in haematology and serum biochemistry profiles were classified using the reference ranges determined by the diagnostic laboratories which performed the tests. The hematologic profiles performed following chemotherapy administration were assessed and recorded for the incidence and grade of neutropenia as defined by VCOG CTCAE 1.1.9 Non-hematologic AEs, such as gastrointestinal upset, were documented by owner-based descriptions and clinician observations during veterinary visits.

Stage and sub-stage were assessed using the World Health Organization clinical staging system. Abdominal ultrasound and bone marrow aspiration were not routinely performed during the study period. Presumption of stage V lymphoma was based on assessment of other clinical abnormalities (such as ocular or pulmonary involvement) or identification of circulating neoplastic cells on blood film evaluation by a board-certified clinical pathologist.

Patients were identified with intent-to-treat with either CHOP19 or CHOP12, with schedules as detailed in Table 1. Patients that had either progressive disease, were euthanized, or received rescue chemotherapy prior to completion of the first cycle were excluded. No maintenance chemotherapy was administered in any patients that attained remission following completion of the CHOP protocol. Patients that were enrolled in a clinical trial involving investigational compounds during the CHOP protocol were excluded.

Chemotherapy was delayed 2 to 7 days if the neutrophil count at nadir was below 1500/μL, if platelet count was less than 50 000/μL, or if owners reported any toxicity that contraindicated treatment. Prophylactic oral enrofloxacin at 5 to 10 mg/kg every 24 hours for 5 days was implemented if the neutrophil count was below 1000/μL, and patients were hospitalized for supportive care if febrile or otherwise ill. Patients with a toxicity that resulted in a delay of a normally-scheduled chemotherapy administration had a 10% to 20% dose reduction of the limiting chemotherapy at the next administration, dependent on clinician discretion.

VCOG criteria10 were utilized to assess response to treatment, and was based upon physical examination at each visit, and additional tests if indicated. CR was defined as the lymph nodes returning to normal size with no other evidence of disease. Partial response (PR) was defined as a 30% decrease in the sum of the target lymph nodes. Progressive disease was considered increase of target lymph node size of 20% or greater beyond smallest recorded measurement, or the development of new lesions. Stable disease was classified as either no change in lymph node size or change of less than 30% decrease or 20% increase, with no new lesions developing. The initial assessment of treatment response was performed at 42 days as is mandated in the VCOG criteria.10

Following the diagnosis of progressive disease of lymphoma, rescue protocols were offered to the owners. Rescue protocols were determined based on attending clinician discretion for each patient. Repeated CHOP protocol was offered if the dog had experienced at least 3 months of remission following completion of the previous CHOP protocol. The success of each protocol and total number of protocols were recorded for each patient. Dogs that had received six doses of DOX were either switched to mitoxantrone for subsequent treatments, or concurrently administered dexrazoxane with DOX after the sixth dose.

Echocardiograms were not routinely performed for patients receiving DOX; dogs that had concern for decreased systolic function at initiation of CHOP protocol and that did not receive any doses of DOX were excluded from the present study. Patients diagnosed with the multi-drug resistance gene (MDR1) mutation were excluded from the present study data analysis because of differences in DI, as routine prophylactic dose reductions of 20% to 40% of VCR and DOX were performed at our institution for patients with MDR1 mutations.

2.3 DI calculations

DI of the cytotoxic chemotherapy drugs in CHOP12 was calculated as described by Longo et al11 in reference to CHOP19 (Table 2). Prednisone was excluded from DI calculations as the prednisone schedule was identical for each protocol with respect to both dose and temporal length of prednisone treatment. The standard planned total administration dose for each cytotoxic chemotherapy drug was calculated and divided by the total number of weeks of the protocol, which provided an mg/m²/week. The DI of each drug in CHOP12 was divided by the DI of each drug in CHOP19 to provide a ratio for relative DI. Overall DI was then calculated by adding the DI for each drug and divided by the total number of drugs. Delivered DI was calculated for each patient population by calculating the actual delivered mg/m² of each chemotherapeutic drug for each week of the CHOP protocol and dividing this by the total number of weeks required for each respective patient to complete their CHOP protocol. Average and relative DI was stated as a mean ± SD.

2.4 Data analysis

Data collected from medical records were contained in a spreadsheet for analysis. Continuous data were expressed as medians, ranges and means ± SD (where appropriate), and categorical data were expressed as frequencies and percentages. PFS, the primary endpoint of this study, was defined as the time interval from date of initiation of chemotherapy treatment to the date of relapse or progression of lymphoma, or death from any cause. OS, a secondary endpoint, was defined as the time from date of initiation of chemotherapy treatment to the date of death from any cause. Dogs that remained alive at the end of the follow-up period or were lost to follow-up were censored at the date of last contact.

Kaplan-Meier estimation was used to demonstrate and display patients' PFS and OS probabilities. The Log-Rank test was used to analyse the PFS and OS between groups. Based on literature review, we anticipated an approximately 8-month PFS in CHOP19 group. True non-inferiority trials generally require very large numbers to be adequately powered12 and, as such, for this exploratory analysis a power calculation was designed to detect a 15% difference in the 8-month progression-free rate between CHOP19 group and CHOP12 group; that is, 50% vs 65% were found to be free from progression at 8 months, respectively. It was felt that a 15% difference in temporal efficacy would be a reasonable limit of clinical relevance. Based on this calculation, 28 dogs in each group would detect a 15% difference with 80% power and a P-value of 0.05. Objective response rate was defined as the combination of patients that experienced complete or partial response. Cox proportional hazards regression was utilized to assess patient characteristics and treatment data in relation to PFS and OS. Categorical data were compared between groups using a two-sided Fisher's exact test, whereas continuous data (eg, age, body weight) was compared between groups using the unpaired two-tailed Students t test after first passing a Kolmogorov-Smirnov normality test. All statistical analyses were performed with a commercial software package (Prism v8.0, Graph-Pad Software, LaJolla, California). A P-value of ≤0.05 was considered statistically significant.

3.1 Patient characteristics

From December 2015 to December 2017, 63 patients were treated with a CHOP protocol; however, only 47 patients were eligible for inclusion in this study. Excluded patients either had primarily extranodal lymphoma (such as hepatosplenic or gastrointestinal forms), severe pre-existing co-morbidity, an MDR1 mutation, or did not receive standard chemotherapy drugs of the CHOP protocol (eg, mitoxantrone instead of DOX). Twenty-eight patients were treated with CHOP12, and 19 patients were treated with CHOP19. Patient characteristics are summarized and statistically compared between groups in Table 3.

Immunophenotype was determined in 40 dogs (85.1%). Seventeen (89.5%) dogs in CHOP19 had immunophenotyping performed, and 4 of 17 (23.5%) were T-cell. Thirteen of 17 dogs had flow cytometry performed (12 B-cell, 1 T-cell), 1 of 17 had PARR performed (1 B cell), and 3 of 17 had ICC performed (1 B-cell, 2 T-cell). Twenty-three (82%) of the CHOP12 dogs had immunophenotyping performed, and 5 of 23 (21.3%) were T-cell. Fifteen of 23 had flow cytometry performed (13 B-cell, 2 T-cell), 2 of 23 had histopathology performed (2 B cell), 5 of 23 had ICC performed (3 B-cell, 2 T-cell), and 1 of 23 had PARR performed (1 T-cell).

Full WHO clinical staging (ie, thoracic and abdominal imaging, bone marrow aspirate) was not routinely performed. In CHOP19 cohort, 100% of the patients had CBC, chemistry panel, and urinalysis performed, 89.5% had chest radiographs performed, and 26.3% had an abdominal ultrasound reported. No patients had abdominal radiographs or bone marrow aspiration performed. All dogs in this study were classified as at least stage III. At presentation, 2 of 19 (10.5%) were considered stage V, and 5 of 19 (26.3%) were classified as sub-stage b. Among the three patients classified as stage V, one had circulating lymphoblasts present on the blood film, one had pulmonary infiltrates, and one had pulmonary and ocular infiltration consistent with lymphoma.

In the CHOP12 cohort, 100% of patients had a CBC, chemistry panel and urinalysis performed, 82% had chest radiographs performed, 7% had an abdominal ultrasound performed, and no patients had a bone marrow aspiration nor abdominal radiographs reported. At presentation, 8 of 28 (28.6%) dogs were considered stage V and 9 of 28 (32.1%) dogs were considered sub-stage b. Of the dogs classified as stage V, four dogs had circulating blasts, two dogs had ocular involvement, one dog had pulmonary infiltrates, and one dog had pulmonary infiltrates with neoplastic pleural effusion.

3.2 Protocols

The 19- and 12-week proposed schedules are detailed in Table 1. None of the dogs in the present study discontinued or changed treatment protocols for reasons other than lymphoma progression. Among dogs in CHOP19, 17 of 19 (89.5%) dogs completed the protocol; two had progressive disease while receiving treatment. Nine of 19 dogs completed the protocol in a 19-week period; average length to completion of CHOP19 was 19.9 weeks, with an SD of 1.3 weeks. Seventeen of 19 dogs had progressive disease by the end of data collection; 13 of 17 (76.5%) of these underwent further treatment with chemotherapy. Overall, 13 dogs in the 19-week group underwent a range of 0 to 3 rescue protocols (median = 1; mean = 1.05). Six of 13 dogs received repeated CHOP19 as a rescue, with mitoxantrone substituted at starting at the second cycle of repeated CHOP, and 7 of 13 dogs received CCNU/L-asparaginase as a first-line rescue therapy. Later rescue protocols included single-agent rabacfosadine (three dogs), CCNU/L-asparaginase (three dogs), single-agent DOX (one dog), and single-agent vinblastine (one dog).

Twenty-one of 28 (75%) dogs completed CHOP12; seven dogs had disease progression while receiving treatment. Six out of 28 dogs completed the protocol as scheduled in a 12-week period; average length to completion of CHOP12 was 13.4 weeks, with an SD of 1.4 weeks. Twenty-seven of 28 dogs had the progressive disease of by the end of the data collection, and 16 of 27 (60%) pursued rescue chemotherapy. Overall, dogs in the 12-week group underwent a range of 0 to 4 protocols (median = 1; mean = 1.14). First line rescue protocols included CHOP12 (2/16) and CCNU/L-asparaginase (14/16), and later rescue protocols included single-agent mitoxantrone (6), single-agent rabacfosadine (4 dogs), single-agent vinblastine (3), CCNU/L-asparaginase (1) and single-agent DOX (1).

The proportion of dogs that received rescue or reinduction chemotherapy after progression was not statistically different between the two protocols (Table 3; P = 0.53). Number of rescue/reinduction protocols per each cohort was compared by the Students t test (P = 0.53) and found not to be statistically different.

3.3 DI calculations and delivery

The targeted DI was calculated for CHOP19 for each drug and the overall protocol and compared with the DI of CHOP12. The DI of CHOP12 was calculated to be 1.32 in comparison to CHOP19, with the VCR DI being lower (0.79) and CTX and DOX DI being higher (1.58 for each). The actual delivered DI was then calculated for each drug in each protocol, and again compared with the target of CHOP19 (Table 2). The overall delivered DI of CHOP19 was 0.930, with the DI for VCR (0.931), CTX (0.932), and DOX (0.928) being lower than the baseline of the calculated CHOP19 DI. The overall delivered DI of CHOP12 was 1.15, with the delivered VCR DI being lower (0.711) and the CTX and DOX delivered DI again being higher (1.40 and 1.35, respectively). The differences in DI for the overall protocol and each individual cytotoxic chemotherapy were found to be statistically significant (P < 0.0001) between CHOP12 and CHOP19 (Table 2).

3.4 Adverse events

There were 143 AEs recorded; 53 for the 19 dogs in CHOP19, and 90 for the 28 dogs in CHOP12. AEs are summarized and graded in Table 4. One of 19 dogs in CHOP19 required hospitalization because of febrile neutropenia following chemotherapy administration; this dog had suspected hepatic dysfunction secondary to lymphoma at the initiation of treatment and despite a lowered dose of VCR (0.4 mg/m²), required hospitalization 6 days following administration. This dog obtained CR and liver values returned to normal; the dog tolerated all future doses of full-dose VCR appropriately. Nine of 19 dogs had dose reductions of chemotherapy; four had DOX dose reduction, and five dogs each had reduction of CTX and VCR. Nine dogs did not have any treatment delays, six had one treatment delay, two had two treatment delays, and two had three delays.

Three of 20 dogs in CHOP12 required hospitalization because of febrile neutropenia following chemotherapy administration; one of these three dogs had received CTX with a referring veterinarian despite displaying a Grade III afebrile neutropenia, and the other two patients were hospitalized following DOX treatment. Eighteen of 28 dogs had dose reductions of chemotherapy; 12 had DOX dose reduction, and eight had reduction of CTX, and six had reduction of VCR. Nine dogs did not have any treatment delays, 10 had one treatment delay, four had two treatment delays, one had three delays, and four had four delays.

3.5 Outcomes

Of the CHOP19 group, 17 of 19 (89.5%) had a complete response, 1 of 19 (5.3%) had a partial response and 1 of 19 (5.3%) failed to achieve a response. The median PFS for the dogs in this group was 245 days (95% confidence interval [CI]: 176-344 days), and the median overall survival (OS) was 351 days (95% CI: 276-489 days). Three dogs remained in remission at the time of study data collection at 322, 422 and 558 days, and two dogs had relapsed but remained alive at the time of study completion. All other dogs in the study died because of lymphoma.

Of the CHOP12 group, 25 of 28 (89.3%) had a complete response, 1 of 28 (3.6%) had a partial response and 2 of 28 (7.1%) failed to achieve a response. The median PFS for the dogs in this group was 141 days (95% CI: 121-192 days), and the median OS was 229 days (95% CI: 163-295 days). One dog died of unrelated causes (biopsy-confirmed osteosarcoma with pathologic fracture) at 571 days and remained in clinical remission for the lymphoma at the time of euthanasia, and one dog came out of remission at 218 days and was lost to follow-up following that time. All other dogs died because of lymphoma.

Comparison of the CHOP19 and CHOP12 PFS and OS are presented in Figures 1 and 2. Dogs receiving CHOP19 had longer PFS (P = 0.047) and OS (P = 0.013) when compared to CHOP12.

Univariate analysis was performed on all patients together to determine potential prognostic factors for PFS and OS, and statistically significant factors are summarized in Table 5. The small sample size for these variables precluded meaningful multivariate analysis.

Performance of T- and B-cell lymphoma in each group were additionally compared individually. The number of T-cell lymphomas in the present study was likely too low (four in CHOP19, five in CHOP12) for meaningful analysis, however, the median PFS of the CHOP12 group was 126 days and median PFS of the CHOP19 group was 159 days (P = 0.140). Among B-cell lymphomas, the CHOP12 group had a median PFS of 168 days and the CHOP19 group had a median PFS of 296 days (P = 0.191).