T staging the partial ground glass nodule

Distinguishing adenocarcinoma in situ from invasive adenocarcinoma has been a topic of discussion within the lung pathology community since the mid-1990’s (1). The fundamental observation at that time was that certain histologic features of small lung adenocarcinomas predicted metastatic potential. Some small peripheral adenocarcinomas completely retain an essentially normal alveolar architecture with no evidence for stromal modification. These tumors have no metastatic potential and are relatively easily recognized as adenocarcinoma in situ. Tumors with this pure form are relatively rare. Others small carcinomas, the commonest group, show at least some obvious desmoplastic fibroblastic response, have clearly established metastatic potential and are readily recognized as invasive adenocarcinomas. The third and most problematic group are those small lung adenocarcinomas associated with bland fibroelastotic stroma that expands alveoli but whose histology is otherwise similar to normal underlying lung architecture. In the original description, these were thought to show “alveolar collapse”. Small tumors with this pattern, like the first group, also rarely if ever metastasize. 

In subsequent years, the additional observation was made that tumors which are mixtures of these various patterns show a metastatic potential which correlates with the invasive component of the tumor and not the overall size. This has led to the adoption by the two main cancer staging agencies, the AJCC, starting in 2018 and the UICC currently, of separate size measurements for the invasive and the non-invasive component (2). Possibly driven by the increasingly narrow bands representing the T substage in the current proposal, the goal of this policy shift was to ensure that patients’ risk for metastasis was not seriously overstated and in the most serious cases, given adjuvant chemotherapy for relatively lower risk disease. 

The challenge has been that the reproducibility of recognizing invasion in these tumors is only fair. Multiple criteria have been investigated to improve this distinction including changes in elastin content/pattern, stromal mucin content, myofibroblastic differentiation and fibroblast proliferation. While all show statistical correlation with this distinction, individual cases may still be problematic.  Hence making a diagnosis of non-invasive adenocarcinoma may be false comfort.

A historical perspective on this topic may be useful. For many years, the fibroelastotic stroma of spiculated lung adenocarcinomas was commonly interpreted as pre-existing scar due to the lack of fibroblastic proliferation typically seen in tumor desmoplasia elsewhere. These tumors were therefore known as “scar carcinoma” with the implication that the tumor derived from regions of pre-existing scar. Only with serial radiologic evaluation and with immunohistochemical evaluation of the stroma was it recognized that the stroma was not pre-existing scar but was response to tumor (3). Since these tumors were well recognized to have significant metastatic potential, there was never any doubt that they represented invasive carcinoma. However, this misunderstanding highlights the way in which routine histologic evaluation of lung carcinoma stroma can be misleading.

Furthermore, it is well recognized in other tumors that different forms of invasion have markedly different biologic impact. In basal cell carcinoma, aggressive vs non-aggressive patterns of invasion are a well-recognized risk factor for recurrence. In colon, budding vs non-budding invasion is well recognized as a risk factor for metastatic disease. Hence, there should be no problem in recognizing that different patterns of invasion may be associated with different metastatic risks in lung cancer.  Lung adenocarcinoma is also somewhat special in that there is not a well-defined pre-existing structure into which the tumor invades. This situation is different from other organs e.g. the tubular GI tract, with pre-existing submucosa and muscular layers which indicate invasion. 

In light of these various considerations, I propose the non-committal term “tumor associated matrix remodeling” for regions in which the adjacent stroma associated with lung carcinomas is distinctly different from that in the adjacent non-neoplastic lung. This term reflects the changes which is encompassed by both “alveolar collapse” and the desmoplastic stromal response in obviously invasive carcinoma while not committing to either in any specific case. It seems likely that the role of the pathologist would then shift from the binary of determining whether there is or is not invasion to the more practical problem of determining absence of invasion vs low grade invasion vs high grade invasion. This allows pathologists to recognize a low but non-zero risk of metastatic disease for some of these tumors as well as retaining a role for a degree of diagnostic uncertainty. It would be worth-while to determine if this new proposal would reduce the palpable discomfort among many pathologists that the AJCC/UICC systems requirement imposes, while improving reproducibility in diagnosis.

(1) Noguchi M, Morikawa A, Kawasaki M et al. Small adenocarcinoma of the lung. Histologic characteristics and prognosis. Cancer. 1995;75:2844-2852.

(2) Detterbeck FC, Boffa DJ, Kim AW, Tanoue LT. The Eighth Edition Lung Cancer Stage Classification. Chest. 2017;151:193-203.

(3) Carter DH, Sloan P, Aaron JE. Immunolocalization of collagen types I and III, tenascin, and fibronectin in intramembranous bone. J Histochem Cytochem. 1991;39:599-606.