The most commonly accepted criteria for spontaneous regression (SR) of cancer were postulated by Everson and Cole in 1959. Here SR is defined as the partial or complete, temporary or permanent, disappearance of the tumor in the absence of any specific therapy [1]. The actual incidence of spontaneous regression is unknown [2]. By the Everson and Cole definition, it was estimated to be not more than 1 in 60,000 to 100,000 cases [3].

Compared with other solid tumors in which SR is more frequently reported, like renal cell carcinoma or malignant melanoma, SR of primary lung cancer is extremely rare [2, 4]. Kumar and colleagues list only two cases of primary lung cancer from 1951 to 2008 that meet the definition of Everson and Cole [5].

There are several suggested mechanisms including immunological factors, hormonal changes, trauma or variations in blood supply [2, 4, 6].

Once regarded as a poorly immunogenic tumor, NSCLC has recently emerged as a target for promising cancer vaccines and immune modulators [7, 8]. Nivolumab, a monoclonal antibody for the human programmed cell death 1 (PD-1) receptor, has been approved this year in the United States for the treatment of squamous NSCLC. In a phase I clinical trial of nivolumab in 2012 33% of the squamous cell lung cancers and 12% of the nonsquamous cell tumors responded with an overall median duration of response of 74 weeks [7]. Interestingly, a smoking history has been associated with an improved response to PD-1 blockade due to a higher neoantigen burden in smokers [9].

Referring to lung cancer, recent articles describe changes in the immunological environment of the tumor that can affect both oncogenesis and regression. Scheider et al. showed an accumulation of regulatory T cells in pulmonary adenocarcinoma and metastatic lymph nodes, resulting in a local decrease of antitumor immune response by natural killer cells [10]. Iwakami and colleagues reported an infiltration of CD8-positive lymphocytes in small cell lung cancer that regressed spontaneously, indicating that T cell-mediated cytotoxity is a possible mechanism of SR in lung cancer [11]. Isobe et al. demonstrated an integrated immune response consisting of immunoglobulin G (IgG) antibodies, CD4 and CD8 T cells against a NY-ESO-1-expressing NSCLC experiencing SR [12].

Several cases of SR have implicated surgery or biopsy conducted on the primary tumor or the metastases as elements that can induce an immunological response [13]. Cole reported that 71 out of 176 cases of SR in cancer were associated with some type of operative trauma [3, 14]. This may be relevant to our patient, who received a biopsy of metastatic lymph nodes before SR was noted. The biopsies were performed with endoscopic forceps, this causing a disruption of the lymph nodes, potentially leading to a release of antigens with activation of the immune system.

There are two important limitations in our case that need to be discussed. In the first place, we were not able to prove malignancy of the tumor in the upper lobe of the right lung. Histological examinations showed extensive necrosis, potentially caused by tumor disintegration. However, being a peripheral lesion, also a lung infarction, an organizing pneumonia or a limited pulmonary vasculitis should also be considered as possible cause. Immunohistochemical analyses to distinguish tumor necrosis from such an infarction, for example by reticulin or keratin staining, were not performed. In conclusion, there is no certain evidence of a primary tumor in this case. Second, by microscopic and immunohistochemical findings, the tumor cells in the paratracheal lymph node were classified as large cell carcinoma. This is rather a diagnosis of exclusion and does not prove a malignancy of the lung exclusively. Carcinomas of several origins and entities show CK7 positivity, so that in theory, this could be a metastasis from elsewhere in combination with a lung infarction.

Summing up our findings, we had a highly suspicious pulmonary lesion with a proven metastatic lymph node in an expectable position for lung cancer. In addition, neither did the staging procedures reveal any other lesions nor did a malignancy develop elsewhere in 7 years of follow-up. Therefore, in our opinion, the most reasonable explanation for the findings in our case is a primary lung cancer with lymph node metastases.