diff --git a/docker/Dockerfile.end-user b/docker/Dockerfile.end-user
index f4ce432037..fcad0a4182 100644
--- a/docker/Dockerfile.end-user
+++ b/docker/Dockerfile.end-user
@@ -34,15 +34,13 @@
 #    echo "FROM dolfinx/dolfinx-onbuild:nightly" | docker build -f- .
 #
 
-ARG PYVISTA_VERSION=0.45.2
-
 # Used to set the correct PYTHONPATH for the real and complex install of
 # DOLFINx
 ARG PYTHON_VERSION=3.12
 # Base image for end-user images
 ARG BASEIMAGE=ghcr.io/fenics/dolfinx/dev-env:current
 
-FROM ${BASEIMAGE} as dolfinx-onbuild
+FROM ${BASEIMAGE} AS dolfinx-onbuild
 LABEL description="DOLFINx in 32-bit real and complex modes (onbuild)"
 
 ARG PYTHON_VERSION
@@ -120,11 +118,11 @@ ONBUILD WORKDIR /root
 
 ########################################
 
-FROM dolfinx-onbuild as intermediate
+FROM dolfinx-onbuild AS intermediate
 
 ########################################
 
-FROM ${BASEIMAGE} as dolfinx
+FROM ${BASEIMAGE} AS dolfinx
 LABEL description="DOLFINx in 32-bit real and complex modes"
 
 ARG PYTHON_VERSION
@@ -149,11 +147,9 @@ ENV PKG_CONFIG_PATH=/usr/local/dolfinx-real/lib/pkgconfig:$PKG_CONFIG_PATH \
 
 ########################################
 
-FROM dolfinx as lab
+FROM dolfinx AS lab
 LABEL description="DOLFINx Jupyter Lab"
 
-ARG PYVISTA_VERSION
-
 WORKDIR /root
 
 RUN pip install --no-cache-dir jupyter jupyterlab
@@ -169,15 +165,9 @@ RUN apt-get -qq update && \
 # matplotlib improves plotting quality with better color maps and
 # properly rendering colorbars.
 # trame is the preferred backend for pyvista.
-RUN dpkgArch="$(dpkg --print-architecture)"; \
-    pip install matplotlib; \
-    case "$dpkgArch" in amd64) \
-      pip install --no-cache-dir pyvista[jupyter]==${PYVISTA_VERSION} ;; \
-    esac; \
-    case "$dpkgArch" in arm64) \
-      pip install --no-cache-dir https://github.com/scientificcomputing/vtk-aarch64/releases/download/v9.4.2-py3.12/vtk-9.4.2.dev0-cp312-cp312-linux_aarch64.whl && \
-      pip install --no-cache-dir pyvista[jupyter]==${PYVISTA_VERSION} ;; \
-    esac; \
+# vtk>=9.5.0 required as minimal version for aarch64 wheels support.
+RUN pip install matplotlib; \
+    pip install --no-cache-dir vtk>=9.5.0 pyvista[jupyter] ; \
     pip cache purge
 
 # Jupyter Notebook kernel specification for complex build DOLFINx