diff --git a/src/infer_request.h b/src/infer_request.h
index cc31a1d3f..bf3610659 100644
--- a/src/infer_request.h
+++ b/src/infer_request.h
@@ -660,13 +660,9 @@ class InferenceRequest {
     secondary_stats_aggregator_ = secondary_stats_aggregator;
   }
 
-  void Cancel()
-  {
-    response_factory_->Cancel();
-    is_cancelled_ = true;
-  }
+  void Cancel() { response_factory_->Cancel(); }
 
-  bool IsCancelled() { return is_cancelled_; }
+  bool IsCancelled() { return response_factory_->IsCancelled(); }
 
 #endif  // TRITON_ENABLE_STATS
 
@@ -774,9 +770,6 @@ class InferenceRequest {
 
   // Sequence I/O states used for implicit state.
   std::shared_ptr<SequenceStates> sequence_states_;
-
-  // Cancellation status
-  std::atomic<bool> is_cancelled_;
 };
 
 std::ostream& operator<<(std::ostream& out, const InferenceRequest& request);