Summary

This study sought to identify and describe characteristics of PTTM with a focus on identifying abnormal clinical characteristics of patients with PTTM and predictors of survival length. We retrospectively identified eight patients with PTTM and demonstrated that it is a deadly complication of adenocarcinoma of various locations and presents with acute right heart failure in patients of every age. All patients described here died within 23 days. In our eight-patient population, BMI, sPAP, d-dimer, NT-proBNP, and CRP were elevated. Low NT-proBNP and sPAP correlated best with survival. Indicators usually used for CAD diagnostics like hsTNT of LVEF were not conclusive as to the length of survival. Low platelets as a manifestation of DIC were only apparent in one of eight patients. Pre-mortem diagnosis of cancer (including diagnosis before the index hospitalization) showed a positive correlation with survival (r = 0.58), but the sample size was not large enough to prove this correlation with Mann–Whitney-U test (p = 0.143). Pre-mortem diagnosis within the index hospitalization, which was exclusive to all patients with CUP, correlated better with survival (Spearman ρ 0.89, log rank 0.014). It is arguable that patients with CUP have a less aggressive form of PTTM or CUP is diagnosed more easily. Initiation of specific therapy was correlated with survival (ρ = 0.786, p = 0.02). Still, because of the emergent referral and oxygen dependency at admission it was an advanced stage of acute PTTM with a poor prognosis in all patients. We also observed a correlation between female sex and longer survival. This may be related to a confounding effect of CUP syndrome mostly being present and diagnosed pre-mortem in our female patients. Our data hints that achieving the diagnosis of adenocarcinoma in PTTM is associated with longer survival, which is more frequently achieved in CUP syndrome, and indicate that affected patients can be identified clinically by signs of right heart failure and metabolic dysregulation.

The clinical view: PTTM is underdiagnosed

The short survival after diagnosis and the high proportion of patients who are not diagnosed with adenocarcinoma during their lifetime demonstrates the diagnostic challenge and consequently the underdiagnosis of adenocarcinoma complicated by PTTM. PTTM is challenging and its suspicion should induce a rapid diagnosis and treatment of cancer. If cancer is not apparent (especially in CUP syndrome) in a conventional scan or with lab values, diagnostic possibilities are very limited and include time-consuming procedures such as dual-energy CT-scan protocols, lung biopsy, and nuclear imaging. While in our study adeno-CUP was the most frequent type of adenocarcinoma, the largest review to date yet described gastric adenocarcinoma as main cancerous entity to develop PTTM. The authors admit that there is much possible bias in that observation [1, 7] especially due to possible underreporting of certain cancer types. Indeed, our study does only show one case of gastric adenocarcinoma. This may indicate that we are underreporting PTTM, and the incidence is higher than described here. PTTM has been described in up to 3% of autopsies of adenocarcinoma patients in general and in one of six patients (17%) with gastric adenocarcinoma in particular [1, 7]. By extrapolation, nine PTTM complications due to gastric adenocarcinoma are likely not to have been diagnosed. However, it is unclear if the undiagnosed cancer patients mostly died of or with PTTM. Only obduction series that clinically correlate right heart failure or right ventricular wall thickness with the occurrence of PTTM can differentiate between correlation and comorbidity. Predominance of gastric cancer in PTTM findings can also be caused by the anatomical proximity (for metastasis) to the lung. Patients with gastric cancer experience excessive weight loss, both due to tumor progression and loss of appetite, possibly leading to slower tumor growth and a longer timeframe for the development of PTTM. In contradiction to CUP syndrome, gastric cancer can more often be an incidental finding of a chest CT-scan, performed because of dyspnea and cough.

In our small population, it remains unclear if CUP syndrome-caused PTTM has a slower progression, or the patients develop the symptoms earlier. This highlights the heterogeneity of PTTM. Still, the data suggests that reaching the diagnosis as it was possible with CUP syndrome improves the survival time of patients. This makes PTTM and its diagnosis an emergency. A common point in the patients we observed is that acute PTTM patients tend to be adipose and to have a diabetes. The median BMI of 28 especially after B-symptoms caused weight loss as well as the median HbA1c of 7.4% indicate that acute PTTM might be linked to an uncontrolled metabolic syndrome. This correlation between cancer and obesity is also known as the so-called obesity paradox in cancer [10] and states that obese patients do have a higher risk of developing cancer, but B-symptoms and cancer-associated appetite loss normalize their weight before initial diagnosis [10]. The authors suggest a different assessment of body weight to prove the correlation to cancer [10]. Another approach could be a measurement of weight loss or weight delta per time unit.

In the following, we would like to further discuss that observation.

The rationale: The Warburg effect

The Warburg effect has been described almost 100 years ago [11, 12] and describes a particular anomalous metabolism of certain cancer cells. They tend to use more glucose than fatty acids to boost their metabolism and can therefore grow faster than other tissues. We have shown that patients with acute PTTM tend to have an uncontrolled metabolic syndrome, possibly associated with type 2 diabetes, and argue that long-time increased sugar levels in these patients lead to an easier metastasis and fuel the development of PTTM. The observation that diabetes increases the risk of cancer has already been shown in various studies and pointed out in consensus papers [12,13,14]. It is also known that insulin increases the risk, at least for certain cancer types [15]. In addition, the incidence of cancer in patients who use biguanides is lowered as demonstrated for liver and colorectal cancer [15], which might also be attributable to the competitive glucose metabolism in normal cells due to increased metabolism and therefore an indirect reduction of the Warburg effect.

We argue that non-adipose patients with PTTM might have a slower progression and have more time for their adenocarcinoma to be diagnosed and treated, so that acute right heart failure mainly puts patients with a fast progression and no cardiac reserve at risk. 

If so, another therapeutic strategy in PTTM might be the use of oral antidiabetics to reduce the glucose excess and slow down tumor metabolism. Because of the low rate of diagnosis, a study including PTTM patients will not be available soon. 

With only three HbA1c values obtained, the quality of Spearman ρ of − 0.89 is biased but a correlation can be suspected.

Lab results and differential diagnosis

Lab results can be misleading in PTTM. The more common diagnosis in known carcinoma associated with right heart failure is PE. A differential diagnosis in patients without adenocarcinoma is chronic thromboembolic pulmonary hypertension (CTEPH). Both entities show higher d-dimer values. Therefore, despite its high sensitivity, d-dimer is a poor predictor of PTTM due to its lack of specificity and its mediocre correlation to the severity of the disease. VEGF as a blood test was shown to have a good correlation [8] with the disease but was not tested in our acute setting. NT-proBNP shows a better correlation, especially in patients with normal LVEF. HsTNT does not correlate with the severity of the disease but still is a useful tool to rule out CAD, which is more commonly diagnosed. In the acute form of PTTM, thrombocytopenia is not common. In our population, one of eight patients had thrombocytes below 150 G/L, which is described in up to 77% of patients [1]. As of now, we cannot explain that high gamma GT values correlate best with the survival length, but we believe this finding to be accidental. Otherwise, a possible explanation is an early liver infiltration caused by the tumor leading to a faster referral to the emergency department. Here sPAP values can be used to assess the patient’s risk. These data show, that laboratory abnormalities caused by PTTM are unspecific, but the combination of the parameters indicating thromboembolic events and heart failure and the exclusion of the differential diagnoses may lead to a suspicion and diagnosis.

While individually nonspecific, the combination of elevated myocardial biomarkers, d-dimers, and HbA1c may represent a diagnostic signature for PTTM in patients presenting with acute right heart failure. Elevated hsTnT and NT-proBNP as well-established markers of myocardial strain and right ventricular dysfunction are commonly seen in PH [16, 17]. A recent study of patients with acute decompensated PH and CTEPH who required intensive care unit admission showed that 57.4% of the patients had increased hsTnT levels and that BNP levels were slightly elevated at 406 (198–623) pg/ml [18]. In our PTTM patient population, the increase of these myocardial biomarkers was more frequent and stronger, with 71.4% having elevated hsTnT and NT-proBNP levels of 7550 (3100–14,000), which underscores the severity of right heart failure in PTTM. D-dimers as a fibrin degradation product are an unspecific marker of macrothrombotic disease and therefore also predictive of venous thromboembolism such as PE [19], a common complication in cancer patients. Thus, they are less useful for differentiating the causes of PH; but, in the absence of PE, they may point towards tumor-associated microthrombosis within the pulmonary vasculature. However, metabolic dysregulation as evidenced by the consistently elevated HbA1c-levels observed in our PTTM patients, even in absence of known diabetes mellitus, is not associated with the risk of venous thromboembolism such as PE [20,21,22]. Since elevated HbA1c-levels are associated with chronic systemic inflammation, endothelial dysfunction, and procoagulant states [23, 24], they may indicate the presence of the paraneoplastic complication of PTTM. Taken together, the combined elevation of hsTnT, NT-proBNP, d-dimers, and HbA1c in patients with unexplained acute right heart failure may serve as a clinical clue for PTTM, and thus should prompt a search for malignancy. Given the high frequency of CUP syndromes in our study, it is open to discussion whether measurement of tumor markers could constitute a valuable additional diagnostic tool. The discovery of this unique biochemical pattern, which can be easily assessed based on blood samples, demonstrates the feasibility of diagnosing PTTM, even in the broad clinical setting of an emergency department.

The biochemical characteristics of patients with PTTM already suggest that this form of PH is particularly severe due to its cancer-associated pathogenesis. This is further supported by the finding that sPAP, which correlates with survival in CTEPH [25], has been reported to be even higher in CTEPH. Recent studies have reported sPAP values between 70 and 90 mmHg for these patients [26, 27], which is higher than the median of 65 mmHg in our population of PTTM patients. While survival is very short overall in PTTM, with the majority of patients dying before receiving specific therapy, and often even before the diagnosis, the prognosis in CTEPH patients has been significantly improved in the recent decades. An analogous Japanese study showed that in CTEPH the proportion of patients receiving specific treatment increased over the last decades to currently 97% [28] and, similarly, the 5-year survival rate rose from 68% at the end of the last century to 93% [28]. Even in a large study of patients with acute decompensation of pulmonary arterial hypertension and CTEPH who required intensive care unit admission, the 28-day survival rate was 60.3% [18]. These data impressively show that PTTM represents a unique pathology and consequently, despite similar clinical presentation, is fundamentally different to this other type of PH. Furthermore, PTTM’s particular dismissal prognosis is emphasized, which may be at least partly due to its diagnostic challenge.

Clinical course

Not only does sPAP correlate with the length of survival, the rise in sPAP can be a useful parameter to assess the progression of the disease [1]. This is confirmed in our population. Patient 7’s sPAP rose by 25 points after two days. In addition, sPAP is also a good long-term parameter. Indeed, in one case, sPAP had been monitored regularly because of a known and controlled class I PH caused by systemic autoimmune collagenosis. SPAP had been stable for almost eight years under PH triple therapy with bosentan, sildenafil, and Iioprost, and was monitored regularly. Unbeknownst to having cancer, the first relevant increase in sPAP was observed eight months prior to the lethal PTTM manifestation. With this case, we observed that PTTM caused lung obstruction can have a course of several months. The benefit of the medication remains unclear, as only singular case reports described an increase in survival with WHO class I PH therapy or other medication in PTTM.

Vasoactive and antiproliferative therapy

Various therapies have been tried in patients that were diagnosed pre-mortem including imatinib, tadalafil, bevacizumab, bosentan, oxaliplatin, epoprostanol, and oral anticoagulation but no regiment has been established yet [1, 29, 30]. The longest reported survival of 15 months was reached with combined chemotherapy (cisplatin/gemcitabine) and reported a regression of PTTM, followed by a lethal relapse [31]. Interestingly, a case with combined PDFG-inhibitor imatinib and VEGF inhibitor bevacizumab showed a similar result also culminating in a survival of twelve months followed by a relapse [29]. This indicates that tumor-adapted chemotherapy protocols may be a possible approach to tackle PTTM. A new approach from Chinese scientists with the new tyrosine kinase inhibitor anlotinib showed promising results as a potent way to reduce PH [32]. These data indicate that PTTM can be effectively treated and thus underline our findings that timely diagnosis is crucial to improve survival. An underestimated method for that purpose is the pulmonary wedge aspiration, an over 30-year-old RHC aspiration technique that can expose cancer cells in pathology. It has shown promising results in case reports [29, 33] and should be made available if unclear PH is diagnosed in RHC.

Clinical implications of diagnosing PTTM

Given the opportunity to start an adapted, specific therapy in PTTM, enabled by the feasibility of achieving a diagnosis during lifetime, we investigated the use and efficacy of the current therapeutic options. Fortunately, specific therapies already seem to be applied in clinical practice. In our population, all patients, who received the diagnosis of cancer in addition to PH, were started with specific therapy, including antineoplastic or antithrombotic therapy. One patient received chemotherapy, one refused it, and one was scheduled for chemotherapy. The fact that this patient died before starting chemotherapy suggests that the low proportion receiving chemotherapy may be due to the short survival time and poor clinical condition of our PTTM patients. Therapeutic anticoagulation was initiated more frequently, suggesting that this may be more practicable in these critically ill patients.

Since in our population, patients who were started with targeted therapy had significantly longer survival times, a clinically relevant difference in a disease with an otherwise rapidly fulminant trajectory becomes evident. As all patients who survived more than one week after admission were started with specific therapy, the importance of rapid initiation becomes clear. The longest overall survival time was achieved in the patient receiving chemotherapy, suggesting high efficacy of antineoplastic treatment. Most of the patients diagnosed received only therapeutic anticoagulation, which was associated with survival above the median, thus indicating a benefit of antithrombotic therapy. Our data of prolonged survival following initiation of therapy underline the importance of diagnosing PTTM and the therapeutic implications of this diagnosis. This is consistent with the literature, with emerging case reports suggesting that targeted therapeutic intervention may meaningfully prolong survival in PTTM [1, 29,30,31].

Limitations

Though there is a bias caused by the small number of patients observed with different cancerous entities in a single center, to our knowledge, this is the largest consecutive monocentric series of patients with an acute presentation of PTTM. Therefore, we can assure that lab values and clinical tests were performed with the same standards. Indeed, we admitted all our patients as emergencies. None were elective deferrals, i.e., after a suspicious lung biopsy. Therefore, we observed a very accelerated course of PTTM, and its progression cannot be compared to patients with a mean life expectancy of 9.5 weeks as described in literature [1].

It is important to note that, due to the exploratory nature of our study, all results have to be interpreted as generating hypothesis rather than definitive conclusions. The small sample size limits the analyses to mainly descriptive statistics, aimed at exploring this currently largely unexplored field, and identifying potential directions for further research. Therefore, it is important to note, that larger, randomized trials are needed to validate our results.