Non-Small Cell Carcinoma-Not Otherwise Specified on Cytology Specimens in Patients with Solitary Pulmonary Lesion: Primary Lung Cancer or Metastatic Cancer?

CONTEXT: Subtyping of solitary pulmonary lesion (SPL) in small amount of cytology specimen using a limited panel of immunohistochemistry (IHC) markers is very important to the correct choice of treatment. This study was performed to categorize non-small cell carcinoma-not otherwise specified (NSCC-NOS) on cytology in patients with SPL, especially with regard to the incidence of metastatic cancer.
MATERIALS AND METHODS: We reviewed 91 cases, in which a precise morphology-based, lineage-specific IHC-aided subtyping was not possible, that qualified as NSCC-NOS on cytology. A stepwise clinical approach and IHC of organ-specific markers was performed on each cell block (CB) to exclude metastasis from extrapulmonary malignancies.
RESULTS: Of the 91 evaluated cases, 65 (71.4%) were diagnosed as non-small cell lung carcinoma (NSCLC)-NOS, 24 (26.4%) were metastatic cancer, and the remaining 2 (2.2%) had undetermined diagnoses. The most frequent primary tumor site was the colorectum (41.7%), followed by breast (20.8%), kidney (8.3%), and then stomach, duodenum, liver, pancreas, gallbladder, prostate, and skin (4.2% each, 1 of 24). Moreover, we found that 7 of the 24 patients with metastatic cancer had a history of extrapulmonary malignancy that was unknown at the time of cytology-based diagnosis.
CONCLUSIONS: These results underscored the need for accurate and stepwise clinical correlation to rule out the possibility of pulmonary metastasis from other sites and appropriate but judicious IHC (i.e., CDX2) on CB for SPL to increase refinement of the cytology diagnosis of NSCC-NOS. Copyright:

INTRODUCTION

A solitary pulmonary lesion (SPL) is defined as a single, rounded radiographic opacity surrounded by lung parenchyma; it may be either primary lung cancer or metastatic cancer.[1] Subtyping of SPL in small amount of specimen using a limited panel of immunohistochemistry (IHC) markers is very important for identifying the correct choice of treatment.[23] However, a substantial proportion of cases show a lack of lineage-specific markers (e.g., TTF-1 and p40), necessitating a diagnosis of non-small cell lung carcinoma (NSCLC)-not otherwise specified (NOS). In this situation, it is recommended that pathologists use the term non-small cell carcinoma (NSCC) rather than NSCLC, because the lack of lineage-specific markers on cytology leaves open the possibility of metastatic cancer.[4]

The aim of this study was to share our institutional experience in categorizing NSCC-NOS on cytology cell block (CB) specimens in patients with SPL, especially with regard to the incidence of metastatic cancer.

SUBJECTS AND METHODS

This was a 7-year single-center retrospective diagnostic analytical study done from January 2013 to December 2019 in a tertiary care center. All subjects provided written informed consent to permit the use of their specimens for research. The study protocol was approved by the Institutional Review Board of Dong-A University Medical Center of Korea (DAUHIRB-19-095).

A cohort of 651 consecutive CT-guided transthoracic fine needle aspiration (TTNA) cytology specimens was collected from 651 Korean patients with SPL on radiological studies according to a previously described definition.[1] We excluded cases which had no tumor cells (n = 30), only necrotic debris or blood (n = 21), and benign lesions (n = 20; 1 pulmonary hamartoma, 14 granulomas, 3 abscesses, 2 aspergillomas). We also excluded patients who had no CB specimen (n = 16) or insufficient atypical cells (n = 8), because ancillary IHC was not applicable to lineage-specific IHC marker analysis. A large series of 556 malignant tumors diagnosed on cytology specimens was selected and further characterized by IHC for TTF-1 and p40, which are used to differentiate adenocarcinoma (ADC) from squamous cell carcinoma (SQC), respectively, according to a previously described algorithm.[45] Normal alveolar epithelium and bronchial epithelium were used as internal controls for TTF-1 and p40, respectively. A total of 460 patients who had definitive diagnoses of ADC (n = 240), SQC (n = 160), small cell carcinoma (n = 54), large cell neuroendocrine carcinoma (n = 3), and sarcoma (n = 3) were excluded from the study. Five cases of metastatic carcinoma (2 esophagus, 2 urinary bladder, and 1 oral cavity) showing positive immunoreactivity for p40 were also excluded from this study [Figure 1].

Figure 1

Selection of cytology specimens with applying the criteria of the 2015 WHO classification with subsequent immunohistochemistry for the present study (TTNA, transthoracic fine needle aspiration; SPL, solitary pulmonary lesion; IHC, immunohistochemistry; CB, cell block; ADC, adenocarcinoma; SQC, squamous cell carcinoma; NSCC-NOS, non-small cell carcinoma-not otherwise specified)

Out of 556 malignant cases, 91, in which a precise morphology-based, lineage-specific IHC-aided subtyping was not possible, qualified as NSCC-NOS, according to the algorithm described in the 2015 WHO classification,[6] and met the inclusion criteria. All slides from these 91 cases were retrieved and reviewed by two experienced cytopathologists (H. W. L and M. S. R), including a lung tumor specialist (M. S. R). Further IHC was performed using organ-specific markers on cytology CB (e.g., estrogen receptor, progesterone receptor, and GATA3 for breast cancer; CK20 and CDX2 for colorectal cancer (CRC); and CD10 and PAX8 for kidney cancer), based on the previously described algorithm.[78] Appropriate tissue controls were used for all markers.[7] IHC was carried out using an autostainer (BenchMark Ultra, Ventana Medical system, USA) according to the manufacturer's instructions. When biopsy or surgical pathology specimens that underwent routine processing were available, cytohistopathology correlation was done. All relevant clinical and radiological details of each patient were taken from medical records.

Finally, all of the cases were categorized into three groups after applying the criteria from the 2015 WHO classification[6] using additional IHC and clinical and radiological findings. Group I consisted of cases that were classified as NSCLC-NOS with null immunophenotype. Group II consisted of cases that were classified as metastatic cancer. Group III consisted of cases with unclear immunophenotype and undetermined diagnosis.

Data analysis was done by arranging cases into columns and is shown as mean (range) for continuous variables and as absolute and relative frequencies for categorical variables.

RESULTS

Ninety-one patients with SPL diagnosed as NSCC-NOS on TTNA cytology specimens were eligible for our study, as shown in Figure 1. The included patients consisted of 63 (69.2%) men and 28 (30.8%) women, with ages ranging from 33 to 94 years (mean age, 66.1 years). The radiologic diameter of the SPL ranged from 1.2 to 8.6 cm, with a mean size of 3.9 cm [Table 1].

Table 1

Patient characteristics according to the Group

Characteristics	Overall (n = 91)	Group I (n = 65)	Group II (n = 24)	Group III (n = 2)
Age (years), mean (range)	66.1 (33 -94)	65.9(41 - 94)	67.0(33 - 87)	60(48 - 72)
Gender
Male	63	48	13	2
Female	28	17	11	0
Size of SPL (cm), mean (range)	3.9 (1.2 -8.6)	4.1 (1.3 -8.6)	3.5 (1.2 - 5.5)	3.6 (2.2 - 5.0)

SPL, solitary pulmonary lesion

When the cases were categorized into three groups, 65 (71.4%) cases were in Group I (NSCLC-NOS), 24 (26.4%) cases were in Group II (metastatic cancer), and the remaining 2 (2.2%) cases were in Group III (undetermined diagnosis). There was no significant association with age, gender, or SPL size according to group [Table 1].

Of the 65 cases in Group I, 10 could be reclassified into favored NSCLC subtypes after subsequent biopsy and 13 could be confirmed with a definitive pathologic diagnosis after surgical resection. Forty-two cases could not be characterized on the basis of morphological and IHC features and clinical history and remained diagnosed as NSCLC-NOS, because no further evaluation was performed. After further evaluation, the number of cases diagnosed as NSCLC-NOS decreased from 65 to 42 [Table 2].

Table 2

Distribution of reclassified histologic subtype of 65 cases in Group I after subsequent evaluation

Subsequent evaluation	Histologic subtype	No. of cases
Biopsy	Favor adenocarcinoma	4
(n = 10)	Favor mucinous adenocarcinoma	3
	Favor squamous cell carcinoma	3
Surgical resection	Pleomorphic carcinoma	3
(n = 13)	Non-keratinizing squamous cell carcinoma	3
	Invasive mucinous adenocarcinoma	2
	Large cell carcinoma	2
	Solid adenocarcinoma	1
	Enteric adenocarcinoma	1
	Basaloid squamous cell carcinoma	1
No further evaluation (n = 42)	NSCLC-NOS	42

NSCLC-NOS, non-small cell lung carcinoma-not otherwise specified

Of the 24 cases in Group II, 17 had a previous history of extrapulmonary malignancy at the time of cytology diagnosis for SPL. In the remaining 7 cases, the primary site of the extrapulmonary malignancy was detected simultaneously (n = 1) or was found by backtracing of the cytological diagnosis and clinical evaluation (n = 6). The most frequent primary tumor site was the colorectum (41.7%, 10 of 24), followed by breast (20.8%, 5 of 24), kidney (8.3%, 2 of 24), and then stomach, duodenum, liver, pancreas, gallbladder, prostate, and skin (amelanotic melanoma) (4.2% each, 1 of 24) [Table 3]. Interestingly, 3 of the 10 CRC patients were diagnosed prior to detection of primary CRC and 1 CRC patient was diagnosed synchronously [Figure 2].

Table 3

Primary site distribution of metastatic cancer of 24 cases in Group II according to the clinical history

Primary tumor site	History of extrapulmonary malignancy		No. of cases (%)
	Previously known (n = 17)	Previously unknown (n = 7)
Colorectum	6	4	10 (41.7)
Breast	5	0	5 (20.8)
Kidney	2	0	2 (8.3)
Stomach	1	0	1 (4.2)
Duodenum	1	0	1 (4.2)
Liver	1	0	1 (4.2)
Pancreas	1	0	1 (4.2)
Gallbladder	0	1	1 (4.2)
Prostate	0	1	1 (4.2)
Skin	0	1	1 (4.2)

Figure 2

(a) Cytology cell block showing cluster of non-small cancer cells in one patient who was diagnosed metastatic adenocarcinoma prior to detection of primary colorectal cancer (H and E, x400). (b) Immunohistochemistry showing positive for CDX2 (IHC, x400). (c) Immunohistochemistry showing negative for p40 (IHC, x400). (d) Immunohistochemistry showing negative for TTF-1 (IHC, x400)

Meanwhile, thirty-three out of the 91 cases studied had a history of extrapulmonary malignancy. Among them, 24 (72.7%) were diagnosed as pulmonary metastasis, whereas 9 (27.3%) had newly diagnosed primary lung cancer.

Two cases in Group III were negative for the lineage-specific markers TTF-1 and p40, but positive for CDX2. Although clinical characteristics and radiological findings were used to estimate the relative probability of extrapulmonary malignancy for these two cases, definitive conclusions were not possible on the basis of these parameters.

DISCUSSION

An SPL may be either primary lung cancer or metastatic cancer,[1] and treatment strategies differ between primary lung cancer and metastatic cancer.[39] The identification of many new histology-driven therapeutic targets over the past decade has resulted in an urgent need for histologic subtyping of lung cancer.[1011] Unfortunately, some diagnoses will still be inconclusive (i.e., NSCLC-NOS), because cytology is the only means available for pathologic diagnosis of SPL in many cases.[212]

Given this situation, IHC is a very valuable tool for separation of SPLs into meaningful therapeutic categories.[3] Therefore, CBs have become the preferred specimens for IHC in the cytological diagnostic process, as they can be handled in the same way as histological specimens.[1314] TTF-1 and p40 are currently the best options to differentiate ADC from SQC histotypes, respectively.[4] The use of these lineage-specific IHC markers increases the refinement of a diagnosis, so that a diagnosis of NSCLC-NOS can be avoided in up to 90% of cases.[515] However, it has been reported that both lineage-specific makers are negative in approximately 15–20% of cases.[314] In line with previous studies, the present study showed that 465 (83.6%) cases out of the initial cohort of 556 consecutive malignant tumors could be characterized on the basis of morphological features and lineage-specific IHC and 91 (16.4%) cases remained NSCC-NOS with double-negative for lineage-specific markers.

NSCC-NOS should be further classified into more specific types, whenever possible. When faced with double-negative cases, pathologists should always consider metastatic cancer.[4] A possible solution is to judiciously select a panel of organ-specific IHC markers, when CB specimens are available.[78] After applying a panel of organ-specific markers and clinical history for 91 patients diagnosed with NSCC-NOS, 24 (26.4%) cases were reclassified as metastatic cancer in the present study.

Interestingly, the most frequent primary tumor site was the colorectum (41.7%), followed by breast (20.8%) and kidney (8.3%). Metastasis from CRC very often takes place in the lungs, and pulmonary metastasis has been detected in 10–22% of all CRC patients.[1617] Although cytological features, such as a massive necrotic background, a palisading appearance of the tumor cells, and elongated tumor cell nuclei and multiple irregularly shaped tumor cell nucleoli, led cytopathologists to detect pulmonary metastasis of CRC,[18] we recommend that CDX2 is preferentially used as the differential IHC marker for determining the origin of extrapulmonary malignancies. A previous study revealed that pulmonary metastases of CRC were positive for CK20 in 83% and CDX2 in 99% of cases.[7] Of note, in enteric or mucinous lung ADCs, lung-specific markers such as surfactant proteins, TTF-1, and napsin A tend to be lost, whereas the enteric-type (colonic) marker CDX2 is aberrantly expressed.[61920] In this study, one enteric ADC and 2 invasive mucinous ADCs, which were negative for TTF-1 but partially positive for CDX2 on cytology CB, were diagnosed as NSCLC-NOS on cytology, although they were confirmed via further surgical resection. Therefore, the differentiation of a primary enteric or mucinous ADC from a metastatic carcinoma originating from the gastrointestinal tract can be challenging. Nevertheless, we recommended that appropriate but judicious IHC for CDX2 for SPL increases refinement of the diagnosis of NSCC-NOS. This point is supported by our study showing that CRC is the most frequent primary site of extrapulmonary malignancy and 40% of pulmonary metastases were diagnosed synchronously or prior to detection of primary CRC.

Moreover, SPLs are frequently detected in patients with a previous history of extrapulmonary malignancy. When cases with a history of extrapulmonary malignancy were considered, Jin et al.[7] reported that 62.9% of SPLs were confirmed as metastatic cancer, whereas 37.1% of SPLs were diagnosed as primary lung cancer. In the present study, we found similar results: among 31 cases with a history of extrapulmonary malignancy, 24 (72.7%) were diagnosed as metastatic cancer, whereas 9 (27.3%) had newly diagnosed primary lung cancer. Of the 24 cases diagnosed as metastatic cancer, 17 had a known history of extrapulmonary malignancy at the time of cytology diagnosis for SPL. In the remaining 7 cases, the primary site of extrapulmonary malignancy was detected simultaneously or was found by backtracing of the cytological diagnosis and clinical evaluation. Because it is not uncommon for a patient who currently has or who previously had extrapulmonary malignancy to develop an SPL, every effort must be made to obtain all available information, that is, previous radiological data and medical history. Furthermore, it should be kept in mind that the extrapulmonary malignancy and the SPL could be either synchronous or metachronous (i.e., the extrapulmonary malignancy occurred before the SPL).

There are some limitations of the present study. First, this study had a retrospective design that used data acquired at a single institute over the course of 7 years; therefore, a selection bias was inevitable. The results must still be confirmed by large-scale prospective analyses in multiple institutions. Second, 42 out of 65 cases diagnosed as NSCLC-NOS (Group I) could not be characterized on the basis of morphological and IHC features and clinical history and remained diagnosed as NSCLC-NOS, because no further evaluation such as biopsy or surgical resection was performed. Even though only one-third of cases of NSCLC-NOS could be further characterized on histological evaluation in this study, NSCLC-NOS can be classified as large cell carcinoma, pleomorphic carcinoma, non-keratinizing squamous cell carcinoma, or solid predominant adenocarcinoma on surgical resection. This is due to casual sampling of undifferentiated tumor areas with no expression of lineage-specific markers and a realistic view of these possibilities is presented here. Our results are similar to those of previously reported research.[2122] We suggest that our cytopathology department has the appropriate diagnostic capabilities to carry out this study. Third, the number of included cases of metastatic cancer was so small (n = 24) that the statistical power to detect differences within this cohort may be limited. However, CRC was the most frequent primary tumor (41.7%), which is why we believe that CDX2 might be the first candidate IHC marker in NSCC-NOS cases that are double-negative for lineage-specific markers. Although the current results may reflect processes in our cytopathology department and not be immediately generalizable, we believe that the stepwise approach outlined above will be helpful in the decision-making process in other institutions. Finally, we have made undetermined diagnoses in two cases that were negative for the lineage-specific markers TTF-1 and p40 but positive for CDX2. Although clinical characteristics and radiological findings were used to estimate the relative probability of extrapulmonary malignancy in these two cases, definitive conclusions were not possible on the basis of these parameters.

In conclusion, this 7-year single-center retrospective diagnostic analysis demonstrated that a significant proportion (26.4%) of cases of SPL on chest radiology that received a diagnosis of NSCC-NOS and that were negative for TTF-1 and p40 on cytology CB specimens were reaffirmed as metastatic cancer after applying a panel of organ-specific markers and clinical history, although the vast majority of our study population with SPL represented primary lung cancer. Moreover, we found that seven out of 24 (29.2%) patients with metastatic cancer had an unknown history of extrapulmonary malignancy at the time of cytological diagnosis of SPL. We also found that CRC was the most frequent primary tumor and, therefore, CDX2 might be the first candidate IHC marker in NSCC-NOS cases that are double-negative for lineage-specific markers. Taken together, the results of the current study underscored the need for accurate and stepwise clinical correlation to rule out the possibility of pulmonary metastasis from other sites and appropriate but judicious IHC (i.e., CDX2) for SPL to increase refinement of the cytological diagnosis of NSCC-NOS. These results can serve as reference data for cytopathologists attempting to distinguish between histological types of NSCC-NOS using ancillary IHC in daily clinical practice.

Financial support and sponsorship

This work was supported by the Dong-A University research fund.

Conflicts of interest

There are no conflicts of interest.