Objectives: To explore the prognostic differences between lung cancer followed by breast cancer (LFB) and breast cancer followed by lung cancer (BFL) and the reasons for the differences. Methods: The database we chose was SEER 18 Regs, from which we retrieved data from patients diagnosed with multiple primary standardized incidence rate (MP-SIR) segments of cancer. Results: A total of 7169 patients were included, of whom 979 were patients with LFB and 6190 were BFL patients. The proportion of small cell lung carcinoma in LFB was 4%, which was significantly lower than that in BFL (p<0.001), while the proportion of carcinoid carcinoma in LFB was significantly higher than that in BFL (p<0.001). Survival analysis of LFB and BFL showed a slightly better prognosis for the former than the latter (HR=0.871 (0.804-0.944)), and the difference was statistically significant (p<0.001). The difference was not statistically significant after adjustment for the pathological type of tumor (HR=0.911 (0.827-1.003), p=0.057). Conclusion: LFB has a worse prognosis than BFL, and this difference is explained by the difference in the ratio of the two pathological components. Keywords: Breast cancer, lung cancer, multiple primary cancer Conclusion Lung cancer is still the leading cause of cancer deaths for tumor patients, and the estimated deaths are 76650 (24%) and 66020 (23%) for males and females, respectively(12). Nevertheless, with the promotion of screening programs and the development of therapeutic treatments, the number of long-term survivors, especially those with early-stage lung cancer, is gradually increasing. However, lung cancer is still more common as a second primary cancer to be studied. Breast cancer patients have a better prognosis and a long survival period. At present, there have been many studies on breast cancer patients with multiple primary tumors. In our study, we screened lung cancer survivors who developed second primary breast cancer and breast cancer patients who developed second primary lung cancer in the SEER database. We compared the prognosis and pathologic types of these two groups. Patients with a second primary cancer usually have a poor prognosis. Female breast cancer patients showed a higher incidence of second primary malignancy, which was associated with poorer prognosis.[13] A cohort study showed that patients diagnosed with second primary lung cancer face favorable lung cancer-specific survival within the early period after diagnosis. In this study, the patients with second primary cancer were at lower risk of lung cancer-specific mortality in the first 5 years (HR, 0.77; 95% CI, 0.76?0.78 at <1 year; HR, 0.87; 95% CI, 0.86-0.89 from 1 to <5 years) but at higher risk thereafter (HR, 1.32; 95% CI, 1.27?1.37 from 5 to 10 years), independent of tumor characteristics or cancer treatment. [14] This result for carcinoid tumors was reversed. The second primary cancer had a worse HR (HR=0.375 p<0.001) than the first primary cancer. Carcinoid tumors are slow-growing malignancies that occur most frequently in the gastrointestinal tract (approximately 74%). They can also be found in the bronchus, ovary, lung, thymus, kidney or thyroid gland.[15] Due to the lower malignancy of this tumor, the first primary carcinoid tumor may survive longer than the first primary breast cancer. Additionally, we cannot neglect the difference in the sample size of the two groups. The limitations of the sample size may have contributed to this result. In our analysis, for breast cancer, the distribution of pathological types was the same for both the first and second primary cancers. This result was consistent with the clinical observations. Adenocarcinoma was the major type of the first and second primary cancer. This may be related to the fact that all included patients were female. Additionally, there was a significant difference in the distribution of pathological characteristics in lung cancer. Small cell cancer accounted for 4% of first primary cancers and 12% of second primary lung cancers. Another different type was carcinoid tumors; this type accounted for 9% of first primary cancers and 2% of second primary cancers. The different sample sizes of the two groups and the degree of partitioning of pathological features may have contributed to this result. Consequently, we analyzed the association between pathological type and latency. In the LFB group, carcinoid tumors showed a significant difference between synchronous (3%) and metachronous (10%) diseases. This means that lung carcinoid tumors, as the first primary cancer, may be more prone to metachronous disease with a longer latency period, which may be associated with a longer survival time for prognosis. In addition, there was no statistically significant difference in either lung or breast cancer within this group. In the BFL group, our study showed that the lobular cancer type had a favorable trend (11%) toward the synchronous situation. In other words, for this subset of patients, the latency was shorter with lobular cancer. This result is similar to that of a prior study showing that synchronous bilateral breast cancer was strongly associated with a lobular phenotype compared to metachronous bilateral breast cancer. [16] As the second primary lung cancer, small cell carcinoma tends to occur more than 6 months after the primary cancer (12% vs 9%), while carcinoid tends to occur less than 6 months after the primary cancer. The reasons for this cannot exclude the difference between the uneven size of cases and the distribution of pathological types in this study. It may also be related to the pathological characteristics of neuroendocrine tumors, which need to be further studied. We compared the outcomes of the two groups. The BFL patients had a worse prognosis than LFB patients (HR=0.871 (0.804-0.944)). We speculate that this difference is most likely due to the difference in the proportions of the pathological components of these two populations. On the one hand, in prognostic comparisons of lung cancer pathology, small cell lung cancer had the poorest prognosis, and carcinoid tumors had the best prognosis; however, the BFL patients had a higher proportion of small cell lung cancer and a lower proportion of carcinoid tumors, and thus, this group had a worse prognosis. On the other hand, after adjustment for pathology, HR=0.911 (0.827-1.003), lost its statistical significance, indicating that the prognoses of BFL and LFB are not significantly different when the pathology types are the same. Thus, BFL has a worse prognosis than LFB, influenced by the proportion of pathology type composition. The prognostic difference between LFB and BFL patients has rarely been reported, and even lung cancer followed by other tumors has received little attention from investigators. Therefore, the molecular mechanism of the difference in prognosis is still unknown and needs to be elucidated by subsequent studies. It is certain that radiotherapy may increase the risk of MPC in breast cancer patients.[17] Some studies have shown that breast exposure to high therapeutic doses may be associated with an excess risk for second cancer induction.[18-20] In terms of the impact of radiation treatment on breast cancer patients, studies conducted more than 20 years ago showed that RT increases the risk of lung cancer in these patients.[21] Consequently, we studied the SIRs of breast cancer patients who received radiotherapy. As shown in Supplementary table 3, patients with left-sided breast cancer had a 16.2% higher risk of developing left-sided lung cancer than right-sided lung cancer after radiotherapy. Subgroup analysis showed that patients diagnosed with breast cancer in 2005-2014 had the highest RR values. The reason for this result may be related to the introduction of IMRT. A study showed that the treatment of primary breast carcinoma with the use of IMRT results in increased probabilities for developing secondary malignancies in the healthy contralateral breast or ipsilateral lung compared to the respective risk for an unexposed population.[22] The period from 1975-1994 was the initiation period of IMRT, with low RR values; 1995-2004 was a transitional period that brought about technical developments, with the technique becoming more widespread over that decade. After 2005, the technique became very common, resulting in significantly higher RR values than in the previous period. However, this paper has several limitations. First, the diagnoses of patients in the SEER database span nearly 40 years, leading to slight differences in anatomical staging criteria. Second, this is a retrospective study, excluding all patients with incomplete survival data, and the grouping is not randomized, resulting in selection bias. Finally, the SEER database registers information mainly on patients in the United States, so a larger study is needed to generalize the findings of this paper. In conclusion, LFB has a worse prognosis than BFL, and this difference is explained by the difference in the ratio of the two pathological components.