diff --git a/deepdiff/search.py b/deepdiff/search.py
index 9b1b11a1..ff37c164 100644
--- a/deepdiff/search.py
+++ b/deepdiff/search.py
@@ -106,7 +106,7 @@ def __init__(self,
 
         self.obj: Any = obj
         self.case_sensitive: bool = case_sensitive if isinstance(item, strings) else True
-        item = item if self.case_sensitive else (item.lower() if isinstance(item, str) else item)
+        item = item if self.case_sensitive or use_regexp else (item.lower() if isinstance(item, str) else item)
         _exclude_exact, self.exclude_glob_paths = separate_wildcard_and_exact_paths(set(exclude_paths) if exclude_paths else None)
         self.exclude_paths: SetOrdered = SetOrdered(_exclude_exact) if _exclude_exact else SetOrdered()
         self.exclude_regex_paths: List[Pattern[str]] = [re.compile(exclude_regex_path) for exclude_regex_path in exclude_regex_paths]
@@ -127,7 +127,7 @@ def __init__(self,
             item = str(item)
         if self.use_regexp:
             try:
-                item = re.compile(item)
+                item = re.compile(item, flags=0 if self.case_sensitive else re.IGNORECASE)
             except TypeError as e:
                 raise TypeError(f"The passed item of {item} is not usable for regex: {e}") from None
         self.strict_checking: bool = strict_checking
@@ -236,7 +236,7 @@ def __search_dict(self,
             parents_ids_added = add_to_frozen_set(parents_ids, item_id)
 
             new_parent = parent_text % (parent, item_key_str)
-            new_parent_cased = new_parent if self.case_sensitive else new_parent.lower()
+            new_parent_cased = new_parent if self.case_sensitive or self.use_regexp else new_parent.lower()
 
             str_item = str(item)
             if (self.match_string and str_item == new_parent_cased) or\
@@ -282,7 +282,7 @@ def __search_iterable(self,
 
     def __search_str(self, obj: Union[str, bytes, memoryview], item: Union[str, bytes, memoryview, Pattern[str]], parent: str) -> None:
         """Compare strings"""
-        obj_text = obj if self.case_sensitive else (obj.lower() if isinstance(obj, str) else obj)
+        obj_text = obj if self.case_sensitive or self.use_regexp else (obj.lower() if isinstance(obj, str) else obj)
 
         is_matched = False
         if self.use_regexp and isinstance(item, type(re.compile(''))):
diff --git a/tests/test_search.py b/tests/test_search.py
index 3984349a..f9cd65f1 100644
--- a/tests/test_search.py
+++ b/tests/test_search.py
@@ -369,6 +369,34 @@ def test_regex_in_string(self):
         result = {"matched_values": {"root"}}
         assert DeepSearch(obj, item, verbose_level=1, use_regexp=True) == result
 
+    @pytest.mark.parametrize('case_sensitive', [False, True])
+    @pytest.mark.parametrize('pattern, matching, nonmatching', [
+        (r'\D+', 'ABC', '123'),
+        (r'\S+', 'ABC', ' '),
+        (r'\W+', '!', 'ABC'),
+        (r'\Boo\B', 'book', 'oo'),
+        (r'\N{LATIN CAPITAL LETTER A}', 'A', 'b'),
+        (r'(?-i:ABC)', 'ABC', 'abc'),
+    ])
+    def test_regex_preserves_pattern_syntax(self, pattern, matching, nonmatching, case_sensitive):
+        result = DeepSearch([matching, nonmatching], pattern, use_regexp=True, case_sensitive=case_sensitive)
+        assert result == {'matched_values': {'root[0]'}}
+
+    @pytest.mark.parametrize('pattern, obj, expected_path', [
+        (r"\['\D+'\]$", {'ABC': 1, '123': 2}, "root['ABC']"),
+        (r"\['(?-i:ABC)'\]$", {'ABC': 1, 'abc': 2}, "root['ABC']"),
+    ])
+    def test_regex_preserves_path_syntax(self, pattern, obj, expected_path):
+        assert DeepSearch(obj, pattern, use_regexp=True) == {'matched_paths': {expected_path}}
+
+    @pytest.mark.parametrize('case_sensitive, expected_paths', [
+        (False, {'root[0]', 'root[1]'}),
+        (True, {'root[0]'}),
+    ])
+    def test_regex_case_sensitivity(self, case_sensitive, expected_paths):
+        result = DeepSearch(['ABC', 'abc', '123'], '[A-Z]+', use_regexp=True, case_sensitive=case_sensitive)
+        assert result == {'matched_values': expected_paths}
+
     def test_regex_does_not_match_the_regex_string_itself(self):
         obj = ["We like python", "but not (?:p|t)ython"]
         item = "(?:p|t)ython"