diff --git a/monai/transforms/utils.py b/monai/transforms/utils.py
index 1d3204b7a0..bb514104ae 100644
--- a/monai/transforms/utils.py
+++ b/monai/transforms/utils.py
@@ -157,7 +157,8 @@ def rescale_array(
     """
     Rescale the values of numpy array `arr` to be from `minv` to `maxv`.
     """
-    dtype_ = dtype or arr.dtype
+    if dtype is not None:
+        arr, *_ = convert_data_type(arr, dtype=dtype)
     mina = arr.min()
     maxa = arr.max()
 
@@ -165,10 +166,7 @@ def rescale_array(
         return arr * minv
 
     norm = (arr - mina) / (maxa - mina)  # normalize the array first
-    arr = (norm * (maxv - minv)) + minv  # rescale by minv and maxv, which is the normalized array by default
-
-    ret, *_ = convert_data_type(arr, dtype=dtype_)
-    return ret
+    return (norm * (maxv - minv)) + minv  # rescale by minv and maxv, which is the normalized array by default
 
 
 def rescale_instance_array(
diff --git a/tests/test_scale_intensity.py b/tests/test_scale_intensity.py
index ddc2fb08e1..cf57b4e3a4 100644
--- a/tests/test_scale_intensity.py
+++ b/tests/test_scale_intensity.py
@@ -35,6 +35,18 @@ def test_factor_scale(self):
             expected = p((self.imt * (1 + 0.1)).astype(np.float32))
             assert_allclose(result, p(expected), rtol=1e-7, atol=0)
 
+    def test_int(self):
+        """integers should be handled by converting them to floats first."""
+        for p in TEST_NDARRAYS:
+            scaler = ScaleIntensity(minv=1.0, maxv=2.0)
+            result = scaler(p(self.imt.astype(int)))
+            _imt = self.imt.astype(int).astype(np.float32)
+            mina = _imt.min()
+            maxa = _imt.max()
+            norm = (_imt - mina) / (maxa - mina)
+            expected = p((norm * (2.0 - 1.0)) + 1.0)
+            assert_allclose(result, expected, type_test=False, rtol=1e-7, atol=0)
+
     def test_channel_wise(self):
         for p in TEST_NDARRAYS:
             scaler = ScaleIntensity(minv=1.0, maxv=2.0, channel_wise=True)