Materials and Methods

Case selection—Medical records of 8 institutions were reviewed to identify dogs with multicentric lymphoma treated with a 6-month CHOP-1 protocol between January 2000 and December 2007. Dogs with lymphoma of the skin or gastrointestinal tract were excluded. Selection criteria included a CHOP-1 protocol with 4 planned treatments with doxorubicin (Appendix), a CR after the CHOP-1 protocol, completion of the 6-month CHOP-1 protocol, CHOP-2 protocol at the time of the first relapse, and adequate follow-up information (ie, response to CHOP-2, date of relapse from CHOP-2, additional chemotherapy protocol given, response to further chemotherapy, and date of death or date lost to follow-up). Dogs receiving l-asparaginase with the CHOP-1 protocol, CHOP-2 protocol, or both were included. The CHOP-2 protocol must have included at least 1 planned treatment with doxorubicin; dogs for which doxorubicin was not included in the planned retreatment protocol were excluded.

Medical record review—Patient records were reviewed to determine signalment, lymphoma characteristics, CHOP-1 and CHOP-2 protocols, and clinical outcomes. For each dog, clinical stage at the time of initiation of the CHOP-1 protocol, determined according to the World Health Organization clinical staging system,³ was recorded. Information on clinical stage at the time of initiation of the CHOP-2 protocol was not recorded because this information was not consistently available. For determination of clinical substage, dogs were classified as substage a if they did not have any clinical signs associated with lymphoma and as substage b if they had clinical signs, including gastrointestinal or respiratory signs, hypercalcemia (ie, serum calcium concentration > 12 mg/dL), fever (ie, rectal temperature > 39.4°C [103°F]), hyphema, or uveitis. Only substage at the time of initiation of the CHOP-2 protocol was recorded. Phenotype information, determined by means of immunohistochemical staining of biopsy specimens and results of a PCR assay, flow cytometry, or both, was recorded for those dogs for which these tests were performed. Information related to the CHOP-1 and CHOP-2 protocols that was recorded included dates when protocols were started, dates when CRs were documented, and dates when relapse occurred. Whether dose reductions were required during the CHOP-1 and CHOP-2 protocols was also recorded. Decreases in chemotherapy drug dosages because of a moderate to severe change in clinical status or adverse gastrointestinal or hematologic effects were categorized as potential dosage modifications. Dosages that were decreased ≥ 20% from the standard dosage for the protocol or compared with a patient's previous dosage were considered to have been modified. Decreases of any chemotherapy drug dosage by < 20% were not considered to be a dosage modification.

Statistical analysis—For the CHOP-2 protocol, CR was defined as the disappearance of all clinical evidence of disease (ie, 100% reduction in the size of all measurable lesions) for ≥ 21 days. Relapse was defined as recurrence of lymph node or organ infiltration after CR. Whether the attending veterinarian continued the CHOP-2 protocol for dogs that only achieved a partial remission (ie, > 50% reduction but < 100% reduction in the size of all measurable lesions) was inconsistent, so for the purposes of this study, a partial response was defined as no response. Also, any response less than partial remission and any response for < 28 days was defined as no response. Dogs that died before their first reevaluation were considered nonresponders. Overall second CR rate was defined as the number of dogs achieving CR divided by the total number of dogs treated. The 95% CI was determined for response rate by use of the method described by Newcombe.⁴

Duration of time to complete the CHOP-1 protocol was defined as the time in days from the first to the last CHOP-1 treatments. First-remission duration was defined as the time in days from when a dog first achieved a CR after initiation of the CHOP-1 protocol until relapse. Duration of time off chemotherapy was defined as the time in days from the last chemotherapy treatment in the CHOP-1 protocol until relapse. Second remission duration was defined as the time in days from when a dog first achieved a CR after initiation of the CHOP-2 protocol until relapse. Median remission duration was calculated by use of the Kaplan-Meier method. Dogs were censored from the remission duration analysis if relapse had not occurred before the end of the study or if the dog was lost to follow-up or died of causes other than lymphoma before relapse. The 95% CI was determined for response duration. Overall survival time was not evaluated because of the confounding influences of euthanasia and owner willingness to pursue other treatments.

Risk factors analyzed to determine their effect on whether dogs achieved a CR with the CHOP-2 protocol and SRD included phenotype (B cell vs T cell), CR before first doxorubicin treatment with the CHOP-1 protocol (yes vs no), dosage reductions with the CHOP-1 protocol (yes vs no), duration of time to complete the CHOP-1 protocol, FRD, duration of time off chemotherapy, body weight at the start of the CHOP-2 protocol, and substage (a vs b) at the start of the CHOP-2 protocol. Categorical variables, including dichotomous variables, were analyzed by use of the χ² test of independence and the Fisher exact test. Distribution of continuous values was evaluated by use of the Shapiro-Wilk test. Values were not normally distributed (all P ≤ 0.005); therefore, the Wilcoxon rank sum test was used to screen continuous data for associations between CR and the CHOP-2 protocol. Remission durations for each potential categorical risk factor were estimated by use of the Kaplan-Meier product-limit method, with the log-rank test for censored data used to compare remission duration curves. The relationship between continuous variables and SRD was assessed by use of the Spearman rank correlation method; variables for which a significant correlation was identified were dichotomized at their median values and reanalyzed by use of the Kaplan-Meier and log-rank tests.

Variables with values of P ≤ 0.05 in univariate analyses were examined by means of multivariable survival analysis. Models were chosen by means of backward elimination, and removed variables were confirmed to be nonsignificant by use of the likelihood-ratio χ² test. The Bonferroni correction was used to correct for multiple testing. To evaluate the combined effects of potential risk factors on SRD, multivariable survival analysis was performed by use of the Cox proportional-hazards regression model.

All statistical analyses were performed with standard software.^f,g For the final analysis, 2-sided values of P ≤ 0.05 were considered significant.

Results

A total of 895 dogs with lymphoma treated with the CHOP-1 protocol at the 8 participating institutions were identified during the retrospective search of medical records. Treatment response information was available for 664 of these dogs, of which 520 (78%) had a CR with the CHOP-1 protocol, 85 (13%) had a partial remission, and 44 (7%) had no response (the remaining 15 [2%] were lost to follow-up). Of the 520 that had a CR with the CHOP-1 protocol, 149 (29%) relapsed during the CHOP-1 protocol, 38 (7%) were lost to follow-up, and 260 (50%) were in remission when they completed the CHOP-1 protocol. Outcome following CR was unknown for the remaining 73 (14%) dogs.

Of the 260 dogs in remission when they completed the CHOP-1 protocol, 78 received alternative protocols following relapse (including lomustine [n = 36]; corticosteroids [16]; doxorubicin [3]; mechlorethamine, vincristine, procarbazine, and prednisone [8]; and other protocols [15]), and 111 underwent the CHOP-2 protocol at the time of relapse. Thirty-two dogs had not relapsed, and 39 dogs were lost to follow-up or received no treatment at the time of relapse. Of the 111 dogs that underwent the CHOP-2 protocol, 16 were lost to follow-up after starting the CHOP-2 protocol. The remaining 95 dogs met all criteria for inclusion in the study.

Median age of the 95 dogs included in the study was 8 years (range, 2 to 16 years), and median body weight was 25 kg (55 lb; range, 2 to 67 kg [4.4 to 147.4 lb]). There were 25 mixed-breed dogs, 10 Golden Retrievers, 5 Scottish Terriers, 5 American Cocker Spaniels, and 5 Shih Tzus. The remaining dogs represented 34 other breeds with ≤ 3 dogs each. There were 58 (61%) male and 37 (39%) female dogs and 40 (42%) dogs with B-cell lymphoma and 6 (6%) with T-cell lymphoma (immunophenotype was not known for the remaining 49 [52%] dogs). Seventy-five (79%) dogs were classified as substage a at the beginning of the CHOP-2 protocol, and 18 (19%) were classified as substage b (substage information was not available for the remaining 2 dogs).

Lymphoma was diagnosed on the basis of results of histologic examination of lymph node or organ biopsy samples in 45 dogs and on the basis of results of cytologic examination in 50 dogs. At the time of initial diagnosis, a CBC and serum biochemistry profile were performed on all dogs. Seventy-nine dogs had thoracic radiography performed at the time of initial staging, 18 had abdominal radiography performed, 76 had abdominal ultrasonography performed, and 57 had bone marrow examined cytologically. According to the World Health Organization staging system, 1 dog had stage II lymphoma, 17 had stage III, 50 had stage IV, and 25 had stage V (stage was unknown for 2 dogs).

Median time to CR with the CHOP-1 protocol was 18 days (range, 3 to 60 days). Seventy-two dogs had CR prior to receiving doxorubicin in the CHOP-1 protocol, and 20 dogs had CR after doxorubicin administration. Timing of CR with regard to doxorubicin administration was unknown for 3 dogs. During treatment with the CHOP-1 protocol, 50 dogs required dosage reductions, including 44 dogs in which vincristine dosage was reduced, 13 in which cyclophosphamide dosage was reduced, and 8 in which doxorubicin dosage was reduced. Median time to complete the CHOP-1 protocol was 178 days (range, 133 to 222 days). Timing of recheck examinations of dogs after completion of the CHOP-1 protocol varied slightly; however, most of the dogs (n = 72) had monthly examinations.

Median FRD with the CHOP-1 protocol was 289 days (range, 150 to 1,457 days). Dogs were off chemotherapy for a median of 133 days (range, 7 to 1,305 days) prior to relapse following the first CR.

Response to the CHOP-2 protocol—Seventy-four of the 95 (78%; 95% CI, 69% to 86%) dogs achieved a second CR with the CHOP-2 protocol. Two dogs achieved transient CRs that persisted for < 21 days and were, therefore, not considered to have developed a CR. Median time to CR with the CHOP-2 protocol was 14 days (range, 4 to 69 days). Fifty-eight dogs achieved CR prior to receiving doxorubicin, and 15 dogs achieved CR after doxorubicin administration. Timing of CR with regard to doxorubicin administration was unknown for 1 dog.

During treatment with the CHOP-2 protocol, 33 of 95 dogs required dosage reductions, including a reduction in vincristine dosage in 32 dogs, cyclophosphamide dosage in 9 dogs, and doxorubicin dosage in 5 dogs. Of the 33 dogs requiring dosage reductions with the CHOP-2 protocol, 26 of these had also required dosage reductions with the CHOP-1 protocol.

All dogs were scheduled to receive at least 1 doxorubicin treatment with the CHOP-2 protocol. Fifty-seven dogs received doxorubicin as the sole anthracycline agent during the CHOP-2 protocol. Twenty-six dogs received at least 1 dose of doxorubicin and then were switched to an alternative drug after the first (n = 14), second (9), or third (3) treatment. Twenty-three dogs were switched to mitoxantrone, 2 dogs were switched to actinomycin-D, and 1 dog was switched to liposomal doxorubicin. Five dogs received mitoxantrone as an alternative to doxorubicin because of presumptive doxorubicin-related cardiotoxicosis. Seven dogs did not receive doxorubicin during the CHOP-2 protocol because of progressive disease requiring abandonment of the CHOP-2 protocol for an alternative protocol (n = 5), owner refusal (1), and death (1).

Twenty-nine dogs completed the CHOP-2 protocol in remission. Ten of these dogs continued to receive cyclophosphamide, vincristine, and prednisone as maintenance chemotherapy. Chemotherapy was discontinued for 16 dogs. Further treatment was unknown for 3 dogs. Four dogs went on to receive a third CHOP protocol at the time of relapse.

Time off chemotherapy following completion of the CHOP-1 protocol was significantly associated with the likelihood of response to CHOP-2. Median duration of time off chemotherapy for dogs that achieved a CR (140 days; range, 7 to 1,305 days) was significantly (P = 0.003) longer than that for nonresponders (83 days; range, 28 to 274 days). Forty-two of the 47 (89%) dogs that were off chemotherapy for ≥ 133 days had a CR, compared with 30 of the 45 (67%) dogs that were off chemotherapy for < 133 days. There was no significant difference between responders and nonresponders with respect to immunophenotype (P = 0.31), CR before first doxorubicin administration with the CHOP-1 protocol (P = 0.55), duration of time to complete the CHOP-1 protocol (P = 0.82), substage at the time of initiation of the CHOP-2 protocol (P = 1.0), or body weight at the start of the CHOP-2 protocol (P = 0.60). When the Bonferroni correction was applied, there was also no difference (P = 0.034) between responders and nonresponders with respect to FRD.

Remission duration with the CHOP-2 protocol and risk factors for relapse—Sixty-nine of the 74 dogs that achieved a CR with the CHOP-2 protocol were included in analyses of remission duration (date when CR was achieved was not known for the remaining 5 dogs). Overall median follow-up time for the 69 dogs that achieved CR was 131 days. Of the 69 dogs that achieved CR with the CHOP-2 protocol, 20 were censored in the remission duration analysis (15 were still in CR at the time of data analysis, and 5 were lost to follow-up). Analysis of censored survival curves indicated median SRD was 159 days (95% CI, 126 to 212 days; Figure 1). Estimated probabilities that dogs would still be in CR 6 months, 1 year, and 2 years after achieving CR were 47%, 31%, and 10%, respectively.

Kaplan-Meier curve of SRD for 69 dogs with lymphoma retreated with a CHOP protocol after relapse following initial treatment with a first-line 6-month CHOP protocol. Median overall SRD was 159 days (95% CI, 126 to 212 days). Vertical marks represent censored data.

Citation: Journal of the American Veterinary Medical Association 238, 4; 10.2460/javma.238.4.501

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Kaplan-Meier curve of SRD for 69 dogs with lymphoma retreated with a CHOP protocol after relapse following initial treatment with a first-line 6-month CHOP protocol. Median overall SRD was 159 days (95% CI, 126 to 212 days). Vertical marks represent censored data.

Citation: Journal of the American Veterinary Medical Association 238, 4; 10.2460/javma.238.4.501

• Download Figure
• Download figure as PowerPoint slide

Kaplan-Meier curve of SRD for 69 dogs with lymphoma retreated with a CHOP protocol after relapse following initial treatment with a first-line 6-month CHOP protocol. Median overall SRD was 159 days (95% CI, 126 to 212 days). Vertical marks represent censored data.

Citation: Journal of the American Veterinary Medical Association 238, 4; 10.2460/javma.238.4.501

• Download Figure
• Download figure as PowerPoint slide

Immunophenotype (log-rank test; P = 0.17), CR before the first doxorubicin treatment in the CHOP-1 protocol (log-rank test; P = 0.76), dose reduction during the CHOP-1 protocol (log-rank test; P = 0.65), duration of time to complete the CHOP-1 protocol (rank sum test; P = 0.82), and body weight (rank sum test; P = 0.12) were not significantly associated with SRD. Risk factors significantly associated with SRD in univariate analyses were FRD (P < 0.001) and duration of time off chemotherapy (P = 0.006). In the multivariable analysis, only FRD was significantly associated with SRD. Dogs with longer FRD in the present study had a longer SRD (rank correlation = 0.30; n = 68; P = 0.014). Dogs that had an FRD longer than the median (289 days) had a significantly (P < 0.001) lower hazard of relapse (hazard ratio, 0.29 [ie, 29% as much risk on any day still at risk]; 95% CI, 0.16 to 0.54) than did dogs with an FRD shorter than the median. Median SRD of the 37 dogs with an FRD ≥ 289 days was 214 days (95% CI, 168 to 491 days), and median SRD of the 31 dogs with FRD < 289 days was 98 days (95% CI, 70 to 144 days; P < 0.001; Figure 2). We suspected an interrelationship involving FRD and duration of time off chemotherapy because the log-likelihood ratio changed significantly between the model with both variables present and the model with only time off chemotherapy present, but not between the model with both variables present and the model with only FRD present. The rank correlation value for the relationship between FRD and time off chemotherapy supported this interrelationship (r_sp = 0.95; n = 90; P < 0.001). The univariable Cox proportional-hazards model indicated that dogs with time off chemotherapy that was longer than the median (133 days) had a 0.45 lower hazard of relapse (ie, 45% lower risk of relapse on any day still at risk; 95% CI, 0.25 to 0.80) than did dogs with a shorter time off chemotherapy (P = 0.006). Median SRD for 39 dogs off chemotherapy for ≥ 133 days was 196 days (95% CI, 140 to 471 days), and median SRD for 30 dogs off chemotherapy for < 133 days was 118 days (95% CI, 70 to 179 days; P = 0.006; Figure 3).

Kaplan-Meier curves of SRD for dogs in Figure 1 classified on the basis of whether FRD was longer (dotted line) or shorter (solid line) than median FRD (289 days). Median SRD for 37 dogs with FRD ≥ 289 days (214 days; 95% CI, 168 to 491 days) was significantly (P < 0.001) longer than median SRD for 31 dogs with FRD < 289 days (98 days; 95% CI, 70 to 144 days). See Figure 1 for key.

Citation: Journal of the American Veterinary Medical Association 238, 4; 10.2460/javma.238.4.501

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Kaplan-Meier curves of SRD for dogs in Figure 1 classified on the basis of whether FRD was longer (dotted line) or shorter (solid line) than median FRD (289 days). Median SRD for 37 dogs with FRD ≥ 289 days (214 days; 95% CI, 168 to 491 days) was significantly (P < 0.001) longer than median SRD for 31 dogs with FRD < 289 days (98 days; 95% CI, 70 to 144 days). See Figure 1 for key.

Citation: Journal of the American Veterinary Medical Association 238, 4; 10.2460/javma.238.4.501

• Download Figure
• Download figure as PowerPoint slide

Kaplan-Meier curves of SRD for dogs in Figure 1 classified on the basis of whether FRD was longer (dotted line) or shorter (solid line) than median FRD (289 days). Median SRD for 37 dogs with FRD ≥ 289 days (214 days; 95% CI, 168 to 491 days) was significantly (P < 0.001) longer than median SRD for 31 dogs with FRD < 289 days (98 days; 95% CI, 70 to 144 days). See Figure 1 for key.

Citation: Journal of the American Veterinary Medical Association 238, 4; 10.2460/javma.238.4.501

• Download Figure
• Download figure as PowerPoint slide

Kaplan-Meier curves of SRD for dogs in Figure 1 classified on the basis of whether time off chemotherapy was longer (dotted line) or shorter (solid line) than median time (133 days). Median SRD for 39 dogs off chemotherapy for ≥ 133 days (196 days; 95% CI, 140 to 471 days) was significantly (P = 0.006) longer than median SRD for 30 dogs off chemotherapy < 133 days (118 days; 95% CI, 70 to 179 days). See Figure 1 for key.

Citation: Journal of the American Veterinary Medical Association 238, 4; 10.2460/javma.238.4.501

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Kaplan-Meier curves of SRD for dogs in Figure 1 classified on the basis of whether time off chemotherapy was longer (dotted line) or shorter (solid line) than median time (133 days). Median SRD for 39 dogs off chemotherapy for ≥ 133 days (196 days; 95% CI, 140 to 471 days) was significantly (P = 0.006) longer than median SRD for 30 dogs off chemotherapy < 133 days (118 days; 95% CI, 70 to 179 days). See Figure 1 for key.

Citation: Journal of the American Veterinary Medical Association 238, 4; 10.2460/javma.238.4.501

• Download Figure
• Download figure as PowerPoint slide

Kaplan-Meier curves of SRD for dogs in Figure 1 classified on the basis of whether time off chemotherapy was longer (dotted line) or shorter (solid line) than median time (133 days). Median SRD for 39 dogs off chemotherapy for ≥ 133 days (196 days; 95% CI, 140 to 471 days) was significantly (P = 0.006) longer than median SRD for 30 dogs off chemotherapy < 133 days (118 days; 95% CI, 70 to 179 days). See Figure 1 for key.

Citation: Journal of the American Veterinary Medical Association 238, 4; 10.2460/javma.238.4.501

• Download Figure
• Download figure as PowerPoint slide

Discussion

In the present study, 78% (74/95) of dogs with multicentric lymphoma achieved CR when retreated with the CHOP protocol after successfully completing an initial 6-month CHOP protocol. Previous studies^1,2 have reported that 92% to 100% of dogs achieved a second CR after receiving CHOP protocols, although numbers of dogs in these studies receiving retreatment were small. Garrett et al¹ reported that 10 of 53 dogs undergoing initial treatment with a 6-month CHOP protocol also underwent the CHOP-2 protocol at the time of relapse, and all 10 achieved CR. Rassnick et al² reported that 11 of the 12 dogs that were retreated with the CHOP protocol achieved CR. Second CR rates of 69% to 75% have been reported^5–7 for dogs with multicentric lymphoma treated with shorter (12- to 20-week) CHOP protocols as well. Importantly, results of the present study may not be applicable to dogs with other less common anatomic forms of lymphomas, such as those involving the gastrointestinal tract, skin, or nervous system.

At the time of relapse from an initial CR, dog owners might be reluctant to retreat with another 6-month chemotherapy protocol. Owner reluctance to retreat at the time of relapse might reflect a perception of a poor prognosis or concerns about the potential for toxicosis or the inconvenience of chemotherapy visits.⁶ Rassnick et al² reported that only 12 of 39 (31%) dogs that relapsed from CHOP chemotherapy underwent retreatment. For the 8 institutions that participated in the present study, 69 of the 260 (27%) dogs were treated with protocols other than the CHOP protocol at the time of relapse from the CHOP-1 protocol remission. The inability to provide information about the expected outcome for dogs retreated with the CHOP protocol might be another reason for the low rate of retreatment among dogs that successfully completed their first protocol. To our knowledge, the present study is the first to examine outcome of dogs retreated with the CHOP protocol and to provide prognostic information that might assist owners when deciding whether to pursue rescue therapy for dogs with lymphoma.

The median FRD for dogs with lymphoma receiving 6-month CHOP protocols is reportedly approximately 9 to 10 months.^1,2 Many dogs relapse when they have been off chemotherapy for several months. Garrett et al¹ speculated that the high second CR rate in their study might have been related to the high doses of chemotherapeutic drugs used in their protocol or the fact that patients were no longer receiving chemotherapy at the time of relapse. Relapse of patients when not receiving treatment might mean that cells that are emerging are still responsive to chemotherapy and that relapse has occurred because of the time since treatment. In contrast, lymphomas in patients that relapse during chemotherapy are drug resistant and, as such, would not be expected to be responsive to retreatment. Results of the present report support this concept. Median time off chemotherapy before relapse was 133 days, and duration of time off chemotherapy was significantly associated with likelihood of response to the CHOP-2 protocol. Overall, 42 of 47 (89%) dogs in the present study that were off chemotherapy for ≥ 133 days achieved CR with the CHOP-2 protocol, whereas only 30 of the 45 (67%) dogs that were off chemotherapy for < 133 days achieved CR. On the other hand, a response rate of 67% for dogs off chemotherapy for < 133 days is higher than rates reported for most rescue protocols for dogs with lymphoma that relapse.^8–11 With this in mind, it still is reasonable to attempt the CHOP-2 protocol prior to scheduling alternative rescue protocols. Response to a CHOP-2 protocol for dogs with short times off chemotherapy may be inferior to times for some rescue protocols, and the long-term outcome for these dogs might be improved by not pursuing the CHOP-2 protocol. However, this was not examined in the present study.

Dogs with longer FRD in the present study also had a longer SRD. Lymphomas in dogs that have a shorter FRD may contain drug-resistant populations of cells that survive initial chemotherapy. Eventually, the drug-resistant population as well as a drug-sensitive population progresses. Retreatment with a CHOP protocol may induce a remission by killing the drug-sensitive cells; however, the drug-resistant cells will ultimately emerge as the dominant cell population, leading to relapse after the CHOP-2 protocol.

An interrelationship between FRD and time off chemotherapy was identified for patients in the present study. We attempted to examine these variables separately because FRD might be affected by the duration of the CHOP-1 protocol. The CHOP-1 protocol involves 6 months of treatment; however, situations such as missed appointments or treatment delays might lead to a longer duration for the protocol. Duration of time to complete the CHOP-1 protocol, however, was not significantly associated with the likelihood of achieving CR with the CHOP-2 protocol or with SRD.

In the present study, body weight and substage (a vs b) at start of the CHOP-2 protocol were examined for prognostic importance and neither was significant. Factors that might have occurred during the CHOP-2 protocol were not evaluated. For example, whether dogs had chemotherapy dose reductions during the CHOP-2 protocol was not examined. Recent studies^12,13 have shown the importance of maximizing chemotherapy dosage for dogs with lymphoma. Similarly, replacing doxorubicin with an alternative drug might have had an impact on SRD because this has been reported¹⁴ to affect FRD. However, factors that were present at the time the CHOP-2 protocol was started were most clinically relevant in helping to predict potential patient response for owners. Future prospective studies to validate the prognostic factors reported here and evaluate others are warranted.

Whether dogs in the present report had dosage reductions in the CHOP-1 protocol was not found to be associated with SRD. A cutoff of 20% was used to define dosage modification, as dosage reductions of 20% are recommended when severe toxicoses develop,¹⁵ and patients that had at least a 20% reduction likely truly needed a dosage modification. Small reductions in chemotherapy dosage may cause large losses in the ability to cure the tumor because of a steep dose-response curve.¹⁶ It is possible that evaluation of all dosage reductions (including those < 20%) might lead to finding a significant association with SRD.

Immunophenotype was not associated with achieving CR with the CHOP-2 protocol or duration of response in the present report. Only 6 dogs were known to have T-cell lymphoma, and phenotype information was missing for most patients in the study. The T-cell phenotype of canine high-grade lymphoma is generally associated with a poor response to chemotherapy.^7,17 It is possible that dogs with T-cell lymphoma that remain in remission with a CHOP-1 protocol for at least 6 months have a more chemosensitive cancer. Further study of this finding with additional patients is warranted.