Double trouble: combined large-cell neuroendocrine and small-cell lung carcinoma.

Lung cancer has seen spectacular progress during the past 10 years. The widespread use of targeted drugs and immune checkpoint inhibitors for metastatic disease, as guided by DNA and RNA next-generation sequencing (NGS) (1), has significantly prolonged survival and facilitated long-term disease control for approximately 30% of patients (2). Nevertheless, certain constellations remain very problematic, among which high-grade neuroendocrine tumors are arguably the most challenging.

In the current issue of Translational Cancer Research, Ai et al. offer insight into today’s worst case scenario: the co-existence of large-cell neuroendocrine (LCNEC) and small-cell lung carcinoma (SCLC) (3), two tumors types with dismal prognosis and a median overall survival (OS) not exceeding 1 year (4,5). One first challenge for combined LCNEC/SCLC is accurate diagnosis, especially in the setting of metastatic disease, because of the limited material available through small biopsies. This is aggravated by the low frequency of LCNEC and SCLC, 15% and 3% among pulmonary malignancies, respectively, while their coexistence is very rare, <1% of lung cancers and only 1/4 of LCNEC (6). Both are characterized by neuroendocrine differentiation, i.e., immunohistochemical expression of CD56, chromogranin A, or synaptophysin, and a high proliferation rate >10 mitoses/2 mm2, so that despite divergent aspects, i.e., large cell size >3 resting lymphocytes, peripheral palisading, rosettes, organoid nesting and trabeculae for LCNEC vs. the typical oat-cell pattern of SCLC (6), their morphologic distinction remains problematic. Several studies have reported considerable interobserver variability with a potential for misclassification in >20% of diagnoses based on small biopsies (7). However, for the patient reported by Ai et al. (3), the diagnosis fulfilled all formal pathologic criteria and can be considered certain, since initial detection of LCNEC was based on ample surgical material obtained through video-assisted thoracoscopy (VATS), while the SCLC component and mixed character of the tumor were confirmed in two different subsequent rebiopsies from the lung and supraclavicular lymph nodes. The mere performance of three longitudinal tissue biopsies during the relatively short disease course of the reported patient is a remarkable feat of the authors, which not only ensured accuracy of a very difficult and infrequent diagnosis, but also reflects extraordinary quality of medical care, because the practicability of repeat biopsies for metastatic lung cancer is only 50% in academic centers, as demonstrated prospectively (8).

A second major problem with high-grade neuroendocrine lung tumors is the paucity of therapeutic options. Actionable mutations, like EGFR mutations or ALK fusions, are exceedingly rare, with a frequency <5% in LCNEC (9) and even lower in SCLC (10). Therefore, routine molecular workup with NGS is not mandatory for these histologies, but should be considered in the special case of a never-, long-time ex-, or light-smoker with <15 pack-years (11), because patients with EGFR, ALK, RET or other druggable alterations can gain many months of survival under tyrosine kinase inhibitors (TKIs), according to several case reports and small retrospective series (4,10,12). Due to his 40-year-long smoking history, the patient reported by Ai et al. definitely did not qualify for this exception and was therefore not tested (3), however, it should be noted that extremely rare cases of oncogene-driven LCNEC and SCLC in smokers do exist in the scientific literature (10,12).

Beyond the perspective of TKI administration, NGS is also important for the molecular typing of LCNEC, which can have therapeutic relevance, as well. Pivotal studies during the last years have shown that LCNEC comprises two different molecular subsets at the genetic level: “NSCLC-like” (aka “type 1”) LCNEC with bi-allelic TP53 and STK11/KEAP1 or KRAS alterations, vs. “SCLC-like” (aka “type 2”) LCNEC enriched for inactivation of TP53 and RB1 (9,13). In addition, it has been reported that “NSCLC-like” LCNEC respond better to platinum-taxane or platinum-gemcitabine doublets compared to the platinum-etoposide chemotherapy usually employed for SCLC (14). In the particular case reported by Ai et al., these considerations are probably not relevant, because the LCNEC component of a combined LCNEC/SCLC tumor is expected to be “SCLC-like” too, as also demonstrated by the only two such cases with published genetic workup (15,16) (), and “SCLC-like” LCNEC derives similar benefit from all aforementioned regimens (14). Another interesting aspect is that pemetrexed, which was administered in the first line by Ai et al., has shown inferior efficacy than other platinum partners in LCNEC (18), but in retrospect this probably did not impact the clinical course of the index patient, either, because he anyway did not response to subsequent taxane-, etoposide- and irinotecan-based combinations, as well (3). Of note, only 3 additional cases of metastatic combined LCNEC/SCLC have been reported in the literature so far () (15-17), which underlines the importance of the report by Ai et al. as precious evidence about a particularly rare and unfavorable tumor type: based on all 4 published cases, primary resistance to routinely available therapies and very short OS emerge as cardinal features of metastatic combined LCNEC/SCLC. Along the same lines, these tumors have a very poor prognosis also in operable stages, for example the OS for patients with resected mixed SCLC tumors was 3 times shorter if the secondary component was LCNEC compared to non-LCNEC alternatives, mainly adenocarcinoma, squamous or adenosquamous tumors, in a large retrospective analysis (19).

Table 1

Published case reports of metastatic combined LCNEC/SCLC in the literature

#	Stage	Genetic alterations	Surgery	RT	Chemotherapy	Best response (stage IV)	OS from stage IV	Ref.
1	II →	TP53, RB1	Y	N	Carbo/pemetrexed (adj)			(15)
	IV (relapse)	SLC17A6 1	N	N	Carbo/etoposide	PD	<3 months
2	IIIb →	n/a	N	CRT	Cis/etoposide (CRT)			(17)
	IIIB →		N	CRT	Cis/paclitaxel (CRT)
	IV (relapse)		N	N	Tem/cap	PR	≈6 months
					Pemetrexed	PD
					Cis/irinotecan	PD
					Etoposide	PD
3	IV	TP53, RB1 2	N	Y	Carbo/etoposide/atezo	PD	>11 months (ongoing4)	(16)
					Irinotecan	PD
					Nivolumab/ipilimumab	PD
		CDK12 p.spl3			Olaparib/paclitaxel (off-label)	PR
4	IV (current case)	n/a	N	N	Carbo/pemetrexed/beva	PD	7 months	(3)
					Carbo/docetaxel	PD
					Carbo/etoposide	PD
					Carbo/irinotecan	PD

1, whole exome sequencing: TP53 p.R273H, SLC17A6 p.W505L, RB1 p.L267X; also MYH8 p.Q1814K and PTPN5 p.M40I mutations of unknown significance. 2, Tempus xT assay; TP53 p.M246V, RB1 copy-number loss; also copy-number gain of MYCL, and ATP7B germline mutation. 3, Guardant360 CDx ctDNA assay; CDK12 c.2109-1G>A; also TP53 p.M246V, PIK3CA amplification, BRAF amplification, and CCNE1 amplification. 4, this patient was primary refractory to multiple chemotherapies and immunotherapies, but was still alive at 11 months due to an exceptional response to off-label olaparib/paclitaxel, presumably facilitated by the presence of a CDK12 splice site mutation (16). Nonetheless, at the time of last follow-up at 11 months the response was already mixed and some lesions had started growing. LCNEC, large-cell neuroendocrine; SCLC, smallcell lung carcinoma; Y, yes; N, no; RT, radiotherapy; cis, cisplatin; carbo, carboplatin; tem/cap, temozolomide/capecitabine; atezo, atezolizumab; beva, bevacizumab; adj, adjuvant chemotherapy; CRT, chemoradiotherapy; PR, partial remission; PD, progressive disease; OS, overall survival; Ref., reference; n/a, not available.

One hope for these patients today is immunotherapy with PD-(L)1 inhibitors, which have demonstrated efficacy and are routinely available after approval by the U.S. Food and Drug Administration (FDA) and the European Medicines Agency (EMA) for both metastatic SCLC and LCNEC. Upfront administration of atezolizumab or durvalumab in combination with platinum-etoposide could prolong the OS of patients with extensive SCLC by approximately 8–10 weeks in two randomized phase 3 trials (5), while an even larger gain of several months has been observed in retrospective series of patients with metastatic LCNEC receiving PD-(L)1 inhibitors alone or in combination with chemotherapy (4,20). The biological rationale for immunotherapy in LCNEC is multifaceted and includes not only the high mutational burden of these tumors, uniquely >10 mut/Mb in median, which is more that of all other lung cancer subtypes (9), but also the tissue upregulation of immune-related pathways and the high blood T-cell reactivity with readily detectable T-cell receptor repertoire alterations in many patients (13,21). An open question is whether administration of nivolumab or any other immune checkpoint inhibitor available in China in 2020 by Ai et al. might have favorably influenced the clinical course of their patient. That being said, the only published combined LCNEC/SCLC under immunotherapy was refractory to atezolizumab, nivolumab and ipilimumab (#3 in ) (16), which demonstrates the current therapeutic cul-de-sac for these tumors and the pressing need for next-generation immunotherapeutics, such as multi-specific antibodies (22) and cell therapies (23).

The OS of the patient diagnosed by Ai et al. with lung, lymph node and bone metastases was very short, measuring 7 months only, despite administration of 4 different therapy lines (3). This stands in good agreement with the findings of a recent large real-world analysis in 191 metastatic LCNEC patients, which showed shorter OS for de novo compared to secondary stage IV tumors, 8.7 vs. 12.6 months in median respectively, and an even worse outcome in case of multiple metastatic sites (4). Of note, patient attrition between successive treatment lines was approximately 50% in this study and underlines the paramount engagement of Ai et al., who managed to administer 4 different chemotherapy lines despite the very aggressive disease course (#4 in ) (3). The lack of any response to routinely available therapies in this and the three other published patients () additionally highlights the need for novel therapeutic strategies in order to improve clinical outcome (24). For example, a refractory patient with inactivating CDK12 mutation (#3 in ) could achieve an exceptional response lasting over 5 months under off-label olaparib/paclitaxel (16).Unfortunately, due to the low frequency of metastatic LCNEC, very few clinical trials are dedicated to this entity (25), which perpetuates lack of evidence, because LCNEC patients are then severely underrepresented in trials of unselected NSCLC. A list of currently active studies explicitly addressing metastatic LCNEC is given in , of which most would accept patients with combined LCNEC/SCLC tumors, as well. Among the trials of , there is only one really novel drug, namely HPN328, which is an anti-CD3/anti-DLL3-directed bispecific antibody (22).

Table 2

Active clinical trials specifically addressing metastatic LCNEC. Most allow for enrollment of patients with combined LCNEC/SCLC, as well, as shown in the rightmost column, while the only trial testing a novel drug is in italic

Clinical trial ID	Phase	Status	Setting	Regimen	Primary endpoint	LCNEC/SCLC eligible (Y/N)
NCT03728361	II	Active, not recruiting	R/R NEC (incl. SCLC, LCNEC)	Nivolumab + Temozolomide	ORR	Y
NCT05126433 (EMERGE-201)	II	Recruiting	R/R LCNEC (and other tumors)	Lurbinectedin	ORR	Y (no explicit exclusion)
NCT05262985	II	Recruiting	advanced LCNEC (1L)	Durvalumab + platinum/etoposide	PFS	Y (no explicit exclusion)
NCT05470595 (ALPINE)	II	Recruiting	Advanced LCNEC	Atezolizumab + platinum/etoposide	OS	Y (if LCNEC ≥50%)
NCT04471727	I/II	Recruiting	R/R SCLC or any tumor with high DLL3 expression1	HPN328 (aDLL3/aCD3 bispecific Ab)	MTD, PK	Y
NCT04079712	II	Active, not recruiting	R/R NEC (incl. LCNEC)	Cabozantinib +Nivo/Ipilimumab	ORR	N (exclusion)
NCT03976518 (CHANCE)	II	Recruiting	R/R NSCLC with rare histology (incl. LCNEC)	Atezolizumab	DCR	Y (no explicit exclusion)
EudraCT 2020-005942-41	II	Recruiting	R/R LCNEC	Durvalumab + platinum/etoposide	OS	Y (no explicit exclusion)

1, LCNEC is also characterized by DLL3 expression in the majority (>70%) of cases. Advanced disease: stage IV or stage III not amenable to definitive chemoradiation; LCNEC, large-cell neuroendocrine; SCLC, smallcell lung carcinoma; NEC, neuroendocrine cancer; R/R, relapsed/refractory; ORR, objective response rate; DCR, disease control rate; PFS, progression-free survival; 1L, first line; Ab, antibody; MTD, maximum tolerated dose; PK, pharmacokinetics; OS, overall survival; Nivo, nivolumab; Y/N, yes/no.

In summary, high-grade neuroendocrine lung tumors represent an unmet need in modern thoracic oncology with almost no benefit from the spectacular advances of the last decade. In particular, the rare coexistence of LCNEC with SCLC is an extremely unfavorable constellation, characterized by primary resistance to routinely available drugs and very short survival, whose management requires expedient combination of clinical skills, advanced molecular profiling, and access to experimental therapeutics.

The article’s supplementary files as