Cervical lesion grading is critical for effective patient management. A three-tier classification (cervical intraepithelial neoplasia [CIN] grade 1, 2 or 3) based on H&E slide review is widely used. However, for reasons of considerable inter-observer variation in CIN grade assignment and for want of a biomarker validating a three-fold stratification, CAP-ASCCP LAST consensus guidelines recommend a two-tier system: low- or high-grade squamous intraepithelial lesions (LSIL or HSIL). In this study, high-risk HPV E6/E7 and p16 mRNA expression patterns in eighty-six CIN lesions were investigated by RNAscope chromogenic in situ hybridization (CISH). Specimens were also screened by immunohistochemistry for p16INK4a (clone E6H4), and by tyramide-based CISH for HPV DNA. HPV genotyping was performed by GP5+/6+ PCR combined with cycle-sequencing. Abundant high-risk HPV RNA CISH signals were detected in 26/32 (81.3%) CIN 1, 22/22 (100%) CIN 2 and in 32/32 (100%) CIN 3 lesions. CIN 1 staining patterns were typified (67.7% specimens) by abundant diffusely staining nuclei in the upper epithelial layers; CIN 2 lesions mostly (66.7%) showed a combination of superficial diffuse-stained nuclei and multiple dot-like nuclear and cytoplasmic signals throughout the epithelium; CIN 3 lesions were characterized (87.5%) by multiple dot-like nuclear and cytoplasmic signals throughout the epithelial thickness and absence/scarcity of diffusely staining nuclei (trend across CIN grades: P<0.0001). These data are consistent with productive phase HPV infections exemplifying CIN 1, transformative phase infections CIN 3, whereas CIN 2 shows both productive and transformative phase elements. Three-tier data correlation was not found for the other assays examined. The dual discernment of diffuse and/or dot-like signals together with the assay's high sensitivity for HPV support the use of HPV E6/E7 RNA CISH as an adjunct test for deciding lesion grade when CIN 2 grading may be beneficial (e.g. among young women) or when 'LSIL vs. HSIL' assignment is equivocal.