SCIeNtIfIC REPORTS | (2018) 8:4470 | DOI:10.1038/s41598-018-22564-7
Tissue Phenomics for prognostic
biomarker discovery in low- and
intermediate-risk prostate cancer
, Maria Athelogou
, Harald Hessel
, Nicolas Brieu
, Mehmet Yigitsoy
, Martin Baatz
, Alexander Buchner
, Christian G. Stief
, Gerd Binnig
, Günter Schmidt
& Ralf Huss
Tissue Phenomics is the discipline of mining tissue images to identify patterns that are related to clinical
outcome providing potential prognostic and predictive value. This involves the discovery process from
assay development, image analysis, and data mining to the nal interpretation and validation of the
ndings. Importantly, this process is not linear but allows backward steps and optimization loops
over multiple sub-processes. We provide a detailed description of the Tissue Phenomics methodology
while exemplifying each step on the application of prostate cancer recurrence prediction. In particular,
we automatically identied tissue-based biomarkers having signicant prognostic value for low- and
intermediate-risk prostate cancer patients (Gleason scores 6–7b) after radical prostatectomy. We found
that promising phenes were related to CD8(+) and CD68(+) cells in the microenvironment of cancerous
glands in combination with the local micro-vascularization. Recurrence prediction based on the selected
phenes yielded accuracies up to 83% thereby clearly outperforming prediction based on the Gleason
score. Moreover, we compared dierent machine learning algorithms to combine the most relevant
phenes resulting in increased accuracies of 88% for tumor progression prediction. These ndings will be
of potential use for future prognostic tests for prostate cancer patients and provide a proof-of-principle
of the Tissue Phenomics approach.
Tissue Phenomics is the systematic discovery of quantitative descriptors for functional, morphological and spatial
patterns in tissue which correlate with disease progression or drug response. It complements genomics which
links genetic information and disease
by adding information on protein expression levels, cellular interactions
and architectures. Using the latest whole slide imaging hardware, advanced multiparametric image analysis, and
big data knowledge discovery, Tissue Phenomics may supersede the manual discovery of novel scoring algorithms
in histopathology. Since the extracted quantitative descriptors can be stored in large databases and cross-linked
with other biomedical knowledge, Tissue Phenomics also stimulates clinical research, improves clinical drug
development and the treatment of patients. Novel image-based biomarkers, referred to as tissue phenes, leverage
pathologists’ scientic and clinical insight into the information contained in tissue. erefore, it has potential to
impact decisively on medicine in general and on oncology in particular.
Examples of novel tissue- and image-based biomarkers have been described in previous work, e.g. refs
While Beck et al.
followed a hypothesis-free approach, for Galon et al.
and Caie et al.
knowledge about a poten-
tial correlation of certain tissue structures and their properties with clinical outcome already existed. However, in
all three cases quantication and statistical evaluation led to surprising results that could not have been achieved
by manual approaches. In Galon et al.
it was known that the immune system responds to tumor tissue, but it was
not anticipated that the immune response could be described by a relative simple mathematical formalism with
a high prognostic value. For Caie et al.
budding tumor cells in the invasive margin of colon cancers were known
to have a meaning for the prognosis of patients (e.g. refs
), but the high prognostic value of big structures of
buds was discovered only by using Tissue Phenomics. In the work of Beck et al.
structural features signicantly
associated with survival in breast cancer were discovered in the stroma close to the tumor by quantication and
data mining in an unbiased data-driven Tissue Phenomics approach.
Definiens AG, Munich, Germany.
Institute for Pathology, Ludwig-Maximilians-University, Munich, Germany.
Department of Urology, Ludwig-Maximilians-University, Munich, Germany.
Present address: Carl Zeiss Meditec
AG, Munich, Germany. Correspondence and requests for materials should be addressed to N.H. (email: nharder@
Received: 11 October 2017
Accepted: 26 February 2018
Published: xx xx xxxx