diff --git a/src/Text/Pandoc/Readers/Docx/Parse.hs b/src/Text/Pandoc/Readers/Docx/Parse.hs
index e83d764fd2e3..4dc9543cfa05 100644
--- a/src/Text/Pandoc/Readers/Docx/Parse.hs
+++ b/src/Text/Pandoc/Readers/Docx/Parse.hs
@@ -148,13 +148,46 @@ mapD f xs =
   in
    concatMapM handler xs
 
+isAltContentRun :: NameSpaces -> Element -> Bool
+isAltContentRun ns element
+  | isElem ns "w" "r" element
+  , Just _altContentElem <- findChildByName ns "mc" "AlternateContent" element
+  = True
+  | otherwise
+  = False
+
+-- Elements such as <w:shape> are not always preferred
+-- to be unwrapped. Only if they are part of an AlternateContent
+-- element, they should be unwrapped.
+-- This strategy prevents VML images breaking.
+unwrapAlternateContentElement :: NameSpaces -> Element -> [Element]
+unwrapAlternateContentElement ns element
+  | isElem ns "mc" "AlternateContent" element
+  || isElem ns "mc" "Fallback" element
+  || isElem ns "w" "pict" element
+  || isElem ns "v" "group" element
+  || isElem ns "v" "rect" element
+  || isElem ns "v" "roundrect" element
+  || isElem ns "v" "shape" element
+  || isElem ns "v" "textbox" element
+  || isElem ns "w" "txbxContent" element
+  = concatMap (unwrapAlternateContentElement ns) (elChildren element)
+  | otherwise
+  = unwrapElement ns element
+
 unwrapElement :: NameSpaces -> Element -> [Element]
 unwrapElement ns element
   | isElem ns "w" "sdt" element
   , Just sdtContent <- findChildByName ns "w" "sdtContent" element
   = concatMap (unwrapElement ns) (elChildren sdtContent)
+  | isElem ns "w" "r" element
+  , Just alternateContentElem <- findChildByName ns "mc" "AlternateContent" element
+  = unwrapAlternateContentElement ns alternateContentElem
   | isElem ns "w" "smartTag" element
   = concatMap (unwrapElement ns) (elChildren element)
+  | isElem ns "w" "p" element
+  , Just (modified, altContentRuns) <- extractChildren element (isAltContentRun ns)
+  = (unwrapElement ns modified) ++ concatMap (unwrapElement ns) altContentRuns
   | otherwise
   = [element{ elContent = concatMap (unwrapContent ns) (elContent element) }]
 
diff --git a/src/Text/Pandoc/Readers/Docx/Util.hs b/src/Text/Pandoc/Readers/Docx/Util.hs
index 0a869bba8691..f373c53304ee 100644
--- a/src/Text/Pandoc/Readers/Docx/Util.hs
+++ b/src/Text/Pandoc/Readers/Docx/Util.hs
@@ -21,12 +21,14 @@ module Text.Pandoc.Readers.Docx.Util (
                                       , findChildrenByName
                                       , findElementByName
                                       , findAttrByName
+                                      , extractChildren
                                       ) where
 
 import qualified Data.Text as T
 import Data.Text (Text)
 import Text.Pandoc.XML.Light
 import qualified Data.Map as M
+import Data.List (partition)
 
 type NameSpaces = M.Map Text Text
 
@@ -67,3 +69,23 @@ findAttrByName :: NameSpaces -> Text -> Text -> Element -> Maybe Text
 findAttrByName ns pref name el =
   let ns' = ns <> elemToNameSpaces el
   in  findAttr (elemName ns' pref name) el
+
+
+-- | Removes child elements that satisfy a given condition.
+-- Returns the modified element and the list of removed children.
+extractChildren :: Element -> (Element -> Bool) -> Maybe (Element, [Element])
+extractChildren el condition
+  | null removedChildren = Nothing  -- No children removed, return Nothing
+  | otherwise = Just (modifiedElement, removedChildren)  -- Children removed, return Just
+  where
+    -- Separate the children based on the condition
+    (removedChildren, keptChildren) = partition condition (onlyElems' $ elContent el)
+
+    -- Helper function to filter only Element types from Content
+    onlyElems' :: [Content] -> [Element]
+    onlyElems' = foldr (\c acc -> case c of
+                                   Elem e -> e : acc
+                                   _      -> acc) []
+
+    -- Reconstruct the element with the kept children
+    modifiedElement = el { elContent = map Elem keptChildren }
\ No newline at end of file
diff --git a/test/Tests/Readers/Docx.hs b/test/Tests/Readers/Docx.hs
index d9935967f6c2..05b69abf618c 100644
--- a/test/Tests/Readers/Docx.hs
+++ b/test/Tests/Readers/Docx.hs
@@ -235,6 +235,10 @@ tests = [ testGroup "document"
             "collapse overlapping targets (anchor spans)"
             "docx/overlapping_targets.docx"
             "docx/overlapping_targets.native"
+          , testCompare
+            "text in shape format"
+            "docx/text_in_shape_format.docx"
+            "docx/text_in_shape_format.native"
           ]
         , testGroup "blocks"
           [ testCompare
diff --git a/test/docx/text_in_shape_format.docx b/test/docx/text_in_shape_format.docx
new file mode 100644
index 000000000000..763e62bf62c9
Binary files /dev/null and b/test/docx/text_in_shape_format.docx differ
diff --git a/test/docx/text_in_shape_format.native b/test/docx/text_in_shape_format.native
new file mode 100644
index 000000000000..09cfa932c61f
--- /dev/null
+++ b/test/docx/text_in_shape_format.native
@@ -0,0 +1,33 @@
+[ Para
+    [ Str "Last"
+    , Space
+    , Str "update:"
+    , Space
+    , Str "May"
+    , Space
+    , Str "1,"
+    , Space
+    , Str "2017"
+    ]
+, Para
+    [ Str "U"
+    , Str "sing"
+    , Space
+    , Str "Microsoft"
+    , Space
+    , Str "Word"
+    , Space
+    , Str "2007/2010"
+    , LineBreak
+    , Str "for"
+    , Space
+    , Str "Writing"
+    , Space
+    , Str "Technical"
+    , Space
+    , Str "Documents"
+    ]
+, Para [ Str "Valter" , Space , Str "Kiisk" ]
+, Para [ Str "Institute" , Space , Str "of" , Space , Str "Physics," , Space , Str "University" , Space , Str "of" , Space , Str "Tartu" ]
+, Para []
+]
\ No newline at end of file