Skip to contents

George Luther Mutter, MD
Pathologist, Brigham and Women's Hospital
Professor of Pathology, Harvard Medical School

Brigham and Women's Hospital
Department of Pathology
75 Francis Street
Boston, MA 02115

Edit Profile

Research Narrative:

My research has resulted in clinically relevant insights into the pathogenesis, histopathologic presentation and clinical course of endometrial cancer and its precursors.  These will be described in turn.

    First we established monoclonal growth and carry-forward of acquired somatic mutations (lineage continuity) as stigmata of premalignant endometrial disease 1.  This required development of new methodologies useful with paraffin embedded human tissues such as PCR-based testing for conservation of novelly altered alleles and nonrandom X chromosome inactivation. The resultant data confirmed our hypothesis that premalignant epithelial lesions in the endometrium are benign neoplasms prone to malignant transformation.  This overturned longstanding dogma that premalignant endometrial lesions were “hyperplastic” field effects.  The next phase of work built upon the practical diagnostic implications of these insights, such as an expectation of localized emergence, rather than gradual transformation of an at-risk field.

Second, we developed revised diagnostic criteria for clinical recognition of premalignant endometrial lesions, which we refer to as Endometrial Intraepithelial Neoplasia (EIN) then validated their outcome predictive value and utility in clinical practice.  Starting with a training set of genetically defined precancers, computerized morphometric analysis of routinely stained tissues was used to discover those histologic features which are characteristic of mutated premalignant clones 2. There were some surprises, including characteristic architectural features of gland crowding in which gland area exceeds that of stroma, and cytologic change relative to the patients native background endometrium which was not always. When these features were combined with minimum size criteria (a minimum disease burden is required for prediction of heightened cancer outcomes) they were clinically validated by showing a 45-fold increased cancer risk when present 3. The clinical diagnostic entity of EIN was introduced to clinical use in 2001, and it now appears in leading gynecologic pathology textbooks, and has become the standard for cancer prevention studies within the multicenter clinical trial cooperative group, the Gynecologic Oncology Group.  I have followed up on these advances with an educational program targeting pathologists which includes a series of postgraduate courses, and a highly popular educational website ( that gets approximately 70,000 hits monthly in 2009.

Third, preclinical phases of endometrial carcinogenesis have been explored with the use of informative disease-specific biomarkers, yielding novel insights into mechanisms of interaction between genetic and nongenetic modifiers of disease risk.  After establishing that most endometrial cancers and examples of EIN lose function of the PTEN tumor suppressor gene, we showed that tiny clusters of morphologically undistinguished somatically acquired PTEN-null (confirmed to be mutant) glands  are present in the normal proliferative endometria of 43% of endogenously cycling premenopausal women 4. These phenotypically normal clones of PTEN mutant cells, that we have designated "latent precancers"  are retained between menstrual cycles, and for years thereafter. This suggests a model of carcinogenesis in which initiation is a common event in normal tissues, with nongenetic modifiers of progression efficiency defining long term cancer risk.   We have confirmed some aspects of this hypothesis through demonstration that nongenetic  cancer risk reducing exposures (intrauterine device use, progesterone exposure, oral contraceptive use) act through destruction of pre-existing subclinical latent precancers in proportion to the reduced cancer risk 5.  This provides a short term response parameter for cancer protective exposures that can be readily queried in normal tissues using simple immunohistochemistry.