56
IASLC ATLAS OF ALK TESTING IN LUNG CANCER
The threshold for a positive ALK FISH result was established on the basis of analysis of histologic samples (a signal pattern typical for ALK rearrangement in at least 15% of cancer cells), but each individual laboratory needs to determine its own threshold for ALK-negative cytologic specimens until consensus recommendations become available. (See Chapter 3 for more information on cutoff values.) Nonoverlapping tumor cells should be selected for ALK FISH, but most three-dimensional clusters remain amenable to ALK analysis, as the detection of rearrangement signals is not hampered by three-dimensionality. For specimens with a low proportion of tumor cells, use of an automated stage guided by appropriate software to reposition the cancer cells increases the precision of analysis and facilitates review of the FISH result.
ALK IHC in Cytology
Immunohistochemistry to detect overexpression of the ALK protein has recently emerged as a valuable method to screen NSCLC for subsequent FISH analysis and for further evaluation of uncertain FISH findings (as described in detail in Chapter 4). As with histologic specimens, ALK IHC is equally promising with cytologic specimens, including cell blocks and conventional or liquid-based cytologic preparations (Moreira 2012, Martinez 2013, Savic 2013, Tanaka 2013). The accuracy of ALK IHC on Papanicolaou-stained cytologic slides has been high, with a sensitivity and specificity of almost 100% compared with ALK FISH (Savic 2013, Tanaka 2013). This accuracy was achieved with the 5A4 antibody (Novocastra, See Chapter 6 ) and an automated immunostainer (Bond-Max; Leica Biosystems) (Figures 3 and 4). Additional studies are needed to validate this antibody on other platforms and to test the performance of a new standardized ALK IHC assay using the D5F3 antibody (Ventana Medical Systems, Inc.) in cytology.
A
B
C
D
Figure 3. IHC on conventional cytologic slides from malignant pleural effusions of pulmonary adenocarcinoma (x?200), showing ALK-negative (3A) and ALK-positive (3B, 3C) tumors, with an ALK-rearranged H2228 cell line used as a positive control (3D). The 5A4 antibody (Novocastra) and an automated immunostainer (Bond-Max; Leica Biosystems) were used, and all results were confirmed by FISH.